European Journal of Homelessness

European Observatory on Homelessness European Observatory on Homelessness

European Journal European Journal of Homelessness

Homelessness of Homelessness The European Journal of Homelessness provides a critical analysis of policy and practice on homelessness in Europe for policy makers, practitioners, researchers and academics. The aim is to stimulate debate on homelessness and housing exclusion at the

European level and to facilitate the development of a stronger Journalof evidential base for policy development and innovation. The journal seeks to give international exposure to significant national, regional and local developments and to provide a forum for comparative analysis of policy and practice in preventing and tackling home-

lessness in Europe. The journal will also assess the lessons for European Europe which can be derived from policy, practice and research from elsewhere.

European Journal of Homelessness is published twice a year by FEANTSA, the European Federation of National Organisations working with the Homeless. An electronic version can be down- loaded from FEANTSA’s website www.feantsaresearch.org.

FEANTSA works with the European Commission, the contracting authority for the four-year partnership agreement under which this publication has received funding.

The information contained in this publication does not necessarily reflect the position or opinion of the European Commission.

ISSN: 2030-2762 (Print) n 2030-3106 (Online) Volume 10, No. 3 _ 2016 10th Anniversary Issue n European Federation of National Associations Working with the Homeless AISBL Fédération Européenne d’Associations Nationales Travaillant avec les Sans-Abri AISBL

194, Chaussée de Louvain n 1210 Brussels n Belgium Tel.: + 32 2 538 66 69 n Fax: + 32 2 539 41 74 Issue Anniversary th research@feantsa.org n www.feantsaresearch.org 10 Volume 10, No. 3 _ 2016 Volume 10, EUROPEAN JOURNAL OF HOMELESSNESS

Journal Philosophy The European Journal of Homelessness provides a critical analysis of policy and practice on homelessness in Europe for policy makers, practitioners, researchers and academics. The aim is to stimulate debate on homelessness and housing exclusion at the European level and to facilitate the development of a stronger evidential base for policy development and innovation. The journal seeks to give international exposure to significant national, regional and local developments and to provide a forum for comparative analysis of policy and practice in preventing and tackling homelessness in Europe. The journal will also assess the lessons for Europe, which can be derived from policy, practice and research from elsewhere.

Editorial Team Joe Doherty, School of Geography and Geosciences, University of St. Andrews, UK (Guest Editor) Eoin O’Sullivan, School of Social Work and Social Policy, University of Dublin, Trinity College, Ireland (Lead Editor) Volker Busch-Geertsema, GISS (Association for Innovative Social Research and Social Planning), Bremen, Germany (Coordinator of European Observatory on Homelessness) Mike Allen, Focus Ireland, Ireland Isabel Baptista, CESIS (Centro de Estudos para a Intervenção Social), Lisbon, Portugal Lars Benjaminsen, Danish National Centre for Social Research, Copenhagen, Denmark Nicholas Pleace, Centre for Housing Policy, University of York, UK Nóra Teller, Metropolitan Research Institute, Budapest, Hungary

Editorial Assistant Kate Waterhouse, School of Social Work and Social Policy, Trinity College Dublin, Ireland Contributors

Mike Allen Focus Ireland Dublin, Republic of Ireland [email protected]

Vassilis Arapoglou Department of Sociology University of Crete, Greece [email protected]

Lars Benjaminsen Danish National Centre for Social Research Copenhagen, Denmark [email protected]

Dennis Culhane School of Social Policy and Practice University of Pennsylvania, USA [email protected]

Boróka Fehér Budapest Social Centre and its Institutions (BMSZKI) Budapest, Hungary [email protected]

Nicolas Herault Melbourne Institute of Applied Economic and Social Research The University of Melbourne, Australia [email protected]

Guy Johnson School of Global, Urban and Social Studies RMIT University, Australia [email protected]

Marcus Knutagård School of Social Work Lund University, Sweden [email protected] Michele Lancione School of Geography and Planning Cardiff University, UK [email protected]

Nicolas Pleace Centre for Housing Policy University of York, UK [email protected]

Nóra Teller Metropolitan Research Institute Budapest, Hungary [email protected] International Advisory Committee of the European Journal of Homelessness Professor Isobel Anderson (University of Stirling), UK Professor Pedro José Cabrera (Comillas Pontifical University of Madrid), Spain Professor Jochen Clasen (University of Edinburgh), UK Professor Dennis P. Culhane (University of Pennsylvania), USA Dr. Pascal De Decker (Hogeschool Gent), Belgium Professor Emeritus Joe Doherty (University of St Andrews), UK Dr. Evelyn Dyb (Norwegian Institute for Urban and Regional Research), Norway Mr. Bill Edgar (European Housing Research Ltd), UK Professor Suzanne Fitzpatrick (Heriot-Watt University, Edinburgh), UK Professor Paul Flatau (Murdoch University), Australia Professor Stephen Gaetz (York University), Canada Professor Susanne Gerull (Alice Salomon Hochschule Berlin), Germany Professor József Hegedüs (Metropolitan Research Institute Budapest), Hungary Professor Claire Lévy-Vroelant (Université Paris 8 -Vincennes – Saint-Denis), France Professor Thomas Maloutas (Harokopio University, Athens), Greece Dr. Magdalena Mostowska (University of Warsaw), Poland Professor Ingrid Sahlin (Lund University), Sweden Professor Marybeth Shinn (Vanderbilt University), USA Dr. Svetlana Stephenson (London Metropolitan University), UK Professor Antonio Tosi (Politecnico University of Milan), Italy Professor Judith Wolf (UMC St Radboud, Nijmegen), The Netherlands

Consultative Committee of the European Journal of Homelessness Christian Perl, Austria Cristina Avonto, Italy Anne Delépine, Belgium Aida Karčiauskienė, Lithuania Đordana Barbaric, Croatia Marco Hoffman, Luxembourg Jiri Ružicka, Czech Republic Jakub Wilczek, Poland Ole Svendsen, Denmark Filipe Miranda, Portugal Juha Kaakinen, Finland Ian Tilling, Romania Olivier Marguery, France Bojan Kuljanac, Slovenia Thomas Specht, Germany Joan Uribe Vilarrodona, Spain Ioanna Pertsinidou, Greece Kjell Larsson, Sweden Peter Bakos, Hungary Rina Beers, The Netherlands Kathleen McKillion, Ireland Robert Aldridge, UK 6 European Journal of Homelessness _ Volume 10, No. 3 _ 2016

Content

Editorial 7

10th Anniversary Issue

Nicolas Pleace Researching Homelessness in Europe: Theoretical Perspectives 19

Lars Benjaminsen and Marcus Knutagård Homelessness Research and Policy Development: Examples from the Nordic Countries 45

Boróka Fehér and Nóra Teller An Emerging Research Strand: Housing Exclusion in Central and South East Europe 67

Vassilis Arapoglou Researching Housing Exclusion and Homelessness in Southern Europe: Learning Through Comparing Cities and Tracking Policies 87

Dennis Culhane The Potential of Linked Administrative Data for Advancing Homelessness Research and Policy 109

Nicolas Herault and Guy Johnson Homelessness in Australia: Service Reform and Research in the 21st Century 127

Mike Allen Homeless Non-Governmental Organisations and the Role of Research 145

Michele Lancione Beyond Homelessness Studies 163 7

Editorial

In September 1985, a seminar was organised by the National Campaign for the Homeless (Ireland) in Cork, Ireland at the request of the Social Affairs Section of the European Commission, to bring together voluntary organisations across the European community, and to devise policies that would eliminate homelessness. One of the recommendations of the seminar was that: “the European Commission fund an association of organisations working with homeless people in the member states so that they may consult regularly on issues affecting homeless people, on methods that will secure improvements in the conditions of homeless people and advise the Commission on policy that will improve the conditions of homeless people.” This recommendation led to the establishment of the Federation Europeenne d’Associations Nationales Travaillant avec les Sans Abris (FEANTSA) in 1989 and in 1991 the European Observatory on Homelessness (EOH). Those who participated in the seminar noted the need to gather information on the following areas to facilitate the development of programmes to deal with homelessness. These included:

• the extent and nature of homelessness in the member states

• the numbers of homeless people, profiled by age and sex

• the precise legal position of homeless people in the member states

• details of existing projects which help homeless people, especially innovative projects undertaken by the government, voluntary or private sectors

• details of existing studies on the problem of homelessness, and

• the causes of homelessness

Between 1991 and 2005 the EOH produced nearly 40 books and reports. They documented the extent of homelessness in Europe, examined issues of measurement and profiled homeless people. Themed volumes on homeless women and youth, immigration, legal systems, welfare regimes and housing regimes were also produced over this period. Thus, the EOH more than fulfilled the mandate set down by those who participated at the 1985 seminar. These various outputs provided increasingly sophisticated analyses of homelessness at the European level. The European Typology of Homelessness and Housing Exclusion (ETHOS) developed by Bill Edgar and Henk Meert, emerged from this research, and this facilitated more rigorous comparative research. 8 European Journal of Homelessness _ Volume 10, No. 3 _ 2016

Following a restructuring of the EOH in 2006, where the practice of having each member state represented by one researcher was replaced with a smaller more focused EOH, the then joint co-ordinators of the EOH, Bill Edgar and Joe Doherty conceived of the European Journal of Homelessness (EJH) as a vehicle for nurturing and disseminating research on homelessness in Europe. Initially envisaged as an annual themed volume, in 2011 it moved to two open editions per annum. Bill Edgar, who edited the first two volumes of the journal, and Joe Doherty stepped down as co-ordinators of the EOH in 2009, having taken on that role in 1997. Volker Busch- Geertsema took on the role of co-ordinator, with Eoin O’Sullivan taking on Editorship of the EJH. In addition to producing the EJH, the EOH, since 2011, also produces an annual comparative research study and commencing in 2006, organises an annual European research conference on homelessness.

In conceiving the EJH, Edgar and Doherty set out a vision for the EJH where the journal would provide ‘a critical analysis of policy and practice on homelessness in Europe for policy makers, practitioners, researchers and academics.’ Thus, from the beginning the EJH aimed not only to publish theoretically and methodologically robust research on homelessness, but also to provide a forum where researchers would interact symbiotically with policy makers and practitioners. This desire to inform policy was set out explicitly whereby the EJH would ‘facilitate the development of a stronger evidential base for policy development and innovation.’ The EJH set out not only to showcase research on homelessness in Europe, but to also learn from significant developments in policy, theory and methodology by publishing contributions by established researchers in North America and the Antipodes. These contributions have proven a source of stimulation for researchers in Europe and inspired research on for example, the costs of homelessness and the duration of shelter use. While European researchers have learned much from research outside of Europe, equally the aforementioned European Typology of Homelessness and Housing Exclusion (ETHOS), the subject of vigorous debate in the pages of the EJH, has had a global impact on defining and measuring homelessness.

Over the 10 years of publication EJH has published 75 peer reviewed articles, 39 reviews and evaluations of local, national and European homelessness strategies and policies, over 60 think pieces (34) and debates (34), 13 research commentaries and 46 book reviews. The range of topics covered are shown in the table below. Some of the highlights include:

– Debates on FEANTSA’s European Typology of Homelessness and Housing Exclusion (ETHOS) developed by Bill Edgar and Henk Meert then members of the European Observatory on Homelessness in 2005 (Volumes 5.2; 6.2; 8.2)

– Examination of the dimensions and development of Housing First (Volumes 5.2; 6.2; 8.1) Editorial 9

– Participation and the issue of agency: the role of homeless people in shaping their own conditions of living and determining their own futures: issues of choice and constraint (Volumes 4; 5.2)

– Evaluation of the jury report following the EU Consensus Conference on Homelessness December, 2010 (Volume 6.2)

– A sustained recurrent interest in issues of governance especially the competing and collaborative roles of the housing market, local, national and supranational state authorities and various representatives of civil society NGO and community based organisations. And in this context, the concept of fluctuating relevance of welfare regimes (Volumes 1; 2; 3; 6.2; 8.1; 8.2; 9.2)

– Social and housing exclusion and the issue of poverty (Volumes 4; 5.1; 7.1; 7.2; 9.1); and the related processes of punitiveness and criminalisation (Volumes 6.2; 7.2; 8.2)

– Quality and effective delivery of homeless services, the focus in the first two volumes, recently revived (Volumes 1; 2; 8.2)

– Evidence and data collection – an abiding theme: ‘point-in-time’ and ‘longitudinal data’ and ‘qualitative’ and ‘quantitative’ approaches (volumes 3; 5.2; 7.1; 8.2)

– Housing and housing rights (Volumes 1, 2, 3, 4, 5.1, 5.2, 9.1, 9.2; 10.1)

– Pathways into, through and out of homelessness (Volumes 2; 3; 5; 6.2)

Topics covered in EJH 2007-2016

TOPICS KEYWORDS Data measurement, collection, definition (ETHOS) Categories of homeless young people, long term, migrants, older people, rough sleepers, refugees, women, families Tenure, accommodation hostels, shelters, private rented sector, homeownership, temporary accommodation, independent living, community provision, rent arrears Services delivery, procurement, quality, standards, costs Homeless voices (as participation, consumer choice, communities, on the streets agents) Welfare & benefits employability, living conditions, labour market, poverty, deprivation, social/ economic /political (multiple) exclusion, demography Media public opinion, stigma NGO & voluntary sector agency, participation, homeless voices, communities Policies & Solutions sustainability, prevention, housing for unusual groups, staircase model, Housing First, housing led, housing ready, staircase 10 European Journal of Homelessness _ Volume 10, No. 3 _ 2016

State & EU governance welfare regimes, national & local planning and housing strategies (urban/regional), EU strategies Criminalisation marginalisation, enforcement, social exclusion, eviction, stigma, victimisation, fear, control, containment, punitiveness, exclusion (public space, train stations) Consensus national, pan-Europe Rights criminal law, human rights, right to housing Approaches to study pathways, poverty, capabilities, participation, medical studies, (theory and method) longitudinal research, social mix, open method of coordination, structure/agency

All EU member states, with the exception of Cyprus and Malta are referenced in the journal as case studies or as comparators. Twenty-two countries are explicitly cited in the titles of articles, reviews or debates. Outside of the EU Norway, USA, Canada, New Zealand, Uruguay and Australia also feature. A key challenge for the next decade of the EJH is to broaden both the geographical coverage and as Michele Lancione in his contribution to this anniversary edition notes, to encourage submissions from the broad ‘social science family’, including anthropology, criminology and geography.

To mark 10 years of publication, an anniversary issue of the EJH was proposed by the EJH editorial board on the theme of ‘Researching Homelessness in Europe’ focusing on the broad areas of theory, methods and policy impact. The special issue was not intended as an ‘audit’ of the EJH, though reference to and evaluation of the journal’s content and impact was encouraged. Rather we looked for an assessment of the ‘state of play’ of homelessness research in Europe and a consideration of its prospective directions of research development. We approached some of the best-known experts on homelessness who have been associated with the journal to submit contributions. While all replied with enthusiasm, several, because of work commitments, were unable to contribute within our relatively short time frame. We are grateful for the eleven authors below for their submissions. We were minimally directive in that while we suggested areas or topics of focus and reminded each of the journal’s philosophy and remit we recognised each as an established authority and we were primarily interested in their interpretations and judgments; we were not therefore prescriptive as to content or coverage.

Synopsis of the articles ‘Researching Homelessness in Europe: Theoretical Perspectives’ by Nicholas Pleace is the first of 8 articles in this anniversary edition. In an exemplary and adroit handling of a copious and complex literature, Pleace presents a forensic analysis of recent theory development in homelessness. He starts with the ‘new orthodoxy’, a perspective that came to the fore during the 1990s, which established that homelessness is the result of the interplay between structural and individual factors. Editorial 11

While recognising this amalgam of the individual and structural as an advance in understanding, Pleace argues that the new orthodoxy has its own problems arising from its ‘inherent vagueness’ and lack of precision especially with regard to how the causal interaction between the personal (individual) and the structural works. The identification of relatively homogeneous subgroups of homeless people and the demarcation of pathways (into, through and out of homelessness) went someway to address these problems, but they in turn are also seen as problematic. In the absence of appropriate European data, subgroup analysis relied too much on taxonomies developed in the USA where data collection was and remains more comprehensive (see Culhane, this issue). As data collection in Europe improves (see Benjaminsen and Knutagård, this issue) subgroup analysis will become more sensitive to European contexts. Pathways analysis served to highlight agency – a recognition of the cognisance of homeless people themselves in having some, albeit restricted, choice in how they conduct their lives. Welcome as agency is, in Pleace’s view, subgroup analysis and pathways research have swung the theory game away from structural explanations towards an undue emphasis on individual experiences. This tendency labelled ‘cultural gravity’, together with ‘assumptive research’ (that is, the assumption that we know what homelessness is), constitute for Pleace the main obstacles to innovative theory development. As a corrective Pleace argues in the final section of his paper for a refurbishment of structural explanations which embrace not just systems failure (support, welfare and housing etc.) but also takes account of the impact of wider structural realities of poverty and inequality. Pleace here prefigures aspects of later commentaries in this issue by Arapoglou and Lancione.

In their article ‘Homelessness Research and Policy Development: Examples from the Nordic Countries’ Lars Benjaminsen and Marcus Knutagård conduct a nuanced examination of the evolution of recent homelessness policy in Finland, Denmark and Sweden. These countries (along with Norway) have an established history of data collection and measurement of homelessness and a tradition of a relatively (compared with much of Europe) close relationship between research and policy development. Benjaminsen and Knutagård’s paper reflects on the way research evidence on Housing First programmes and interventions in North America has informed the development of experimental Housing First programmes in Nordic countries. The impact has been uneven. In Finland and then Denmark Housing First programmes have been espoused, challenging hitherto dominant ‘treatment first/ housing ready’ approaches. In Sweden, however, the authors suggest that the continuing commitment to a long-established staircase model and the difficulties in accessing affordable social housing have inhibited the adoption of Housing First. As Benjaminsen and Knutagård observe, the interplay between research and policy can be difficult to articulate when policy development does not always follow a 12 European Journal of Homelessness _ Volume 10, No. 3 _ 2016

linear path and research is not necessarily primarily aimed at contributing to new policies and practices. Even when research evidence points to the benefits of new developments, implementation and structural barriers can inhibit their adoption.

Boróka Fehér and Nóra Teller’s ‘An Emerging Research Strand: Housing Exclusion in Central and South East Europe’ recounts a rather different story from that of the Nordic countries and starkly illustrates the uneven relationship across European regions between research and homelessness policy. This is an ambitious paper seeking to ‘summarize the state of the art of research’ in Central Eastern and South Eastern Europe. Citing evidence, particularly from Poland, Slovakia, the Czech Republic, the Baltic states, Romania, Bulgaria, Serbia, Croatia and Slovenia, the authors identify the differential impact of a series of political, economic and social upheavals – the post-Soviet transition to a market economy, the Balkan wars and the recent refugee/immigrant crisis – in exacerbating problems of social exclusion and the challenge of homelessness. There is little or no history in homelessness research in this region. This has led to the instigation of policies based on weak or no evidence, the most alarming example of which is the re-criminalisation of homelessness in some counties – a policy with clear political motivations. However, while conscious of considerable uneven development from country to country, Fehér and Teller highlight the emergence in the past decade of significant advances in the development of housing and welfare research with exploratory qualitative and quantitative studies on the nature and causes of homelessness, an emerging understanding of pathways into and out of homelessness and analyses of the effectiveness of policy responses for selected groups. Mirroring developments in West Europe and overseas there has been a reorientation of research towards a more structural understanding of homelessness. The authors end with a plea for more comparative, interdisciplinary and applied research within and across CEE/ SEE countries.

Set within the framework of recent theoretical developments which have challenged orthodox accounts of policy diffusion and transfer, Vassilis Arapoglou’s paper ‘Researching Housing Exclusion and Homelessness in Southern Europe: Learning Through Comparing Cities and Tracking Policies’ examines the issue of ‘policy mobility’ – the transfer of policy initiatives from one social, economic and political context to others with different societal complexions. The attendant challenges of ‘policy translation’ are scrutinized conceptually and variously illustrated by reference to supported housing, affordable housing and the housing of recently arrived immigrants and refugees. Arapoglou’s account demonstrates that the process of translation is always complex and frequently contested. In particular, Arapoglou highlights the challenges posed in southern Europe by the neoliberal conventions, with attendant austerity tropes, that characterises many homelessness strategies and poverty policies originating in the EU and elsewhere. These he Editorial 13

suggests clash with the distinguishing characteristics of the Mediterranean welfare regime which include civic pride, pluralism, associationalism, local democracy movements, informal and spontaneous acts of solidarity and organised forms of community care and, more recently, a variety of grassroots initiatives stimulated by anti-austerity politics. Referencing Gramsci’s ‘theory of hegemony’ and ‘philosophy of praxis’, Arapoglou examines the contradictions and tensions embedded in policy mobility, and discusses how some of these tensions may be addressed by a considered process of ‘translation’ and a ‘politics of learning’. His conclusion considers the potential of policy mobility approaches for comparing homelessness initiatives within and across different types of welfare states.

Dennis Culhane, in his lucid commentary on ‘The Potential of Linked Administrative Data for Advancing Homelessness Research and Policy’, identifies and explicates the research potential in linking administrative data across multiple social systems and geographical scales. In the context of a widening use of integrated administrative data systems in North America, Europe and Australia, Culhane enumerates the benefits that the use of such data brings to social research: the data has already been collected and does not have to be generated by expensive and time consuming trials and is therefore relatively cost effective, it does not rely on self-reporting and is population-based and, further, as data accumulates over time the potential for longitudinal analysis grows. With regard specifically to homelessness research integrated administrative data offers a low cost, continuous source of measurement of the prevalence and duration of homelessness in a community. When linked with other records, such data can facilitate interrogation of, for example, discharge practices from social welfare systems such as prisons and other institutions in assessing the impact on homelessness. Similarly, interventions to reduce rates of homelessness or to expedite exit from homelessness can be tracked and the impact on other health and social welfare systems evaluated. Notwithstanding Culhane’s enthusiasm for the collation and interrogation of administrative data – an enthusiasm clearly demonstrated over two decades of innovative research only a fraction of which is cited in this article – he is acutely aware of the risks associated with the use of integrated administrative data for social research. The final part of his article is devoted to encouraging an explicit and transparent discussion of the ethical considerations relating to issues of intentionality, privacy and security. In Culhane’s view compilation and analysis of integrated administrative data enterprise requires a new set of legal, ethical, technological and procedural standards. He concludes that the basic ethical requirement for the operation of integrated data systems is a transparent communication strategy that brings together government data owners and community partners with ‘general citizenry (including sceptics)’. In this way, he argues people can participate and ensure appropriate policies for the beneficial use of data. 14 European Journal of Homelessness _ Volume 10, No. 3 _ 2016

Nicolas Herault and Guy Johnson’s ‘Homelessness in Australia: Service Reform and Research in the 21st Century’ begins in 2008 when the newly elected Australian Labour Government made homelessness its highest policy priority. This fundamental change was predicated on the emergence over some years of the so-called ‘new homeless’ – young people and families replacing ‘skid row’ populations – and changes in the housing market characterised by a crisis of affordability and contraction in the availability of social housing. From the beginning research evidence was explicitly identified as key in determining the direction of homelessness initiatives. Overall A$11 million was allocated to support the development of programmes to both prevent homelessness and end chronic homelessness. A$5 million of this total was awarded to Melbourne University to undertake a large scale, national longitudinal study focusing on housing instability and homelessness. The resulting project, ‘Journeys Home’, is the main subject matter of Herault and Johnson’s article. Their informative and detailed account suggests that the research design and methodology of the longitudinal study might serve as a base model for similar studies elsewhere. The finding from analyses of the six-wave dataset cover among other topics health, psychological distress and crime, exit rates and substance usage; these have made important contributions to the evidence base certainly in Australia and, as Herault and Johnson suggest, perhaps internationally – the interrogation of the database continues. Yet, disappointingly, despite ‘Journeys Home’ being recognised around the world as an exceptional dataset capable of answering fundamental questions about the dynamics of homelessness, it has yet to have a meaningful impact on Australian homelessness policy and service delivery. The authors attribute this failure to resistance from entrenched interests committed to transitional support tied to short and medium term accommodation, and to a lack of political leadership and commitment. For Herault and Johnson a key message of the Australian experience is that it is important to have independent non-aligned people and institutions driving policy and practice change and conclude that the bigger lesson is that ‘without structural reform increasing the supply of affordable housing, the capacity of systems reform to reduce homelessness, let alone end it, is limited’.

In the penultimate contribution to this anniversary issue, ‘Homeless Non-Governmental Organisations and the Role of Research’, Mike Allen explores the reciprocal and often complex relationships between homelessness research and NGOs. Allen’s principal focus is on the role and motives of homeless NGOs in commissioning and undertaking research and the impact this research has on NGO practice. In this later context – citing the example of Housing First and the challenge to traditional homelessness taxonomies by Dennis Culhane and his colleagues in North America – Allen also references the influence on NGOs of research emanating from the broader research community. Throughout Allen uses ‘Focus Ireland’, Editorial 15

where he is Director of Advocacy, Communication and Research, as a case study. NGO research reflects all aspects of their multiple functions: evaluations of the process, impact and outcomes of service delivery, assessment of social return on investment, advocacy appraisal and fundraising. But can on occasion also be agenda setting, exampled in this paper by the adoption of regular measurement of homelessness by Irish statutory bodies. The ethical challenges of NGO homelessness research reflect those of the wider research community whether academic or state sponsored, with additional unease regarding the potential exploitation of NGO clients. Citing such concerns, some NGOs refuse to engage in research arguing that their primary role is to address clients’ personal problems; thus, Allen argues, buttressing individual interpretations of homelessness with a consequent downplaying of structural issues. Here Allen cites Amartya Sen’s critical take on ‘positional objectivity’ in support of his arguments, and echoes – though without direct reference – some of the concerns raised by Pleace’s notions of ‘cultural gravity’ and ‘assumptive research’ (see this issue). Addressing ethical concerns is central to Allen’s advocacy of research and Focus Ireland has established a ‘research ethics committee’ to this end. The impact of research findings on the practice of service delivery, Allen suggests, is rarely direct and sometimes hard to detect. Field workers – and some mangers – rarely read research reports. Allen argues for the establishment of conduits for the communication and transfer of research findings. In this context, he cites Focus Ireland’s ‘lunchtime and occasional evening talks’ in partnership with researchers from Trinity College and plans to stream these to other Focus Ireland centres in other Irish cities. Allen also advocates collaborative NGO research, to avoid duplication certainly, but more proactively to enable better resourced, larger studies with robust research designs. Allen concludes his article with an evocation for NGOs to engage in a ‘critical and constructive dialogue with frontline staff and service users’ to tease out and articulate ‘timely questions’ for future research endeavour.

This anniversary edition concludes with Michele Lancione’s article ‘Beyond Homelessness Studies’. This is a challenging and provocative paper calling for a period of self-reflection and auto-critique on the part of the homelessness research community. In the context of major political and social changes evidenced by growing nationalisms, the treat of EU dismemberment, refugee influx, continuing austerity and economic uncertainties, the author argues that it is time to rethink our notions of what homelessness research is all about – to go ‘beyond homelessness study’, to challenge the status quo and to question habitual ways of thinking. Lancione approvingly references Pleace’s identification of ‘cultural gravity’ and ‘assumptive theory’ (see Pleace in this issue) as among the main risks facing current homelessness research and adds a third – which he suggests is implicit in Pleace’s critique – the fact that most research is policy driven, related to the evaluation of 16 European Journal of Homelessness _ Volume 10, No. 3 _ 2016

this or that program, often at a very small scale and with little interaction with broader debates and agendas. Being self-referential it is often ‘out-of-sync’ with the latest advances in social theory, and as a consequence not able to contribute meaningfully to those debates. This is not a blanket criticism. Lancione recognises in 10 years of publication, the EJH has frequently reflected on and posed such questions and that even among the work he castigates there is frequently an implicit and sometimes explicit, appreciation of such problems. Reflecting on some of the papers in this anniversary issue Lancione cites approvingly Culhane’s recognition of the dangers of uncritical data collection and measurement as an objective means to bolster political ambitions, and Herault and Johnson’s acknowledgment that data, no matter how well constructed, can be ‘porous’ because its impact and effectiveness depends on factors that transcend data itself – in the Australian case the opposition of policy makers and the inertia of homelessness practitioners. And he is appreciative of Benjaminsen and Knutagård’s demonstration that ‘context matters’ and how structural forces and political orientations can lead to very different results in homelessness policies and practices. Yet, as an incitement and challenge to homelessness research and homelessness researchers, Lancione concludes with the question: how can this field of study address and overcome the future risks of being assumptive, relativist, policy-driven, self-referential, inertial and potentially positivist?

Finally, we would like to thank all those who have so generously contributed to the journal over the past ten years with a special thanks to the contributors to this anniversary edition for their enthusiasm and for so willingly committing time to writing their articles. We hope this anniversary issue will spark significant response and initiate lively debate which will feature in future issues of the journal.

Joe Doherty and Eoin O’Sullivan, December 2016. 17

10th Anniversary Issue

10th Anniversary Issue 19

Researching Homelessness in Europe: Theoretical Perspectives Nicholas Pleace

Center for Housing Policy, University of York, UK

>> Abstract_ This paper explores the theoretical developments in homelessness research and relates them to the European context. The paper asserts that European academics have relied too heavily on American research and evidence, and that there are dangers in using these ideas and data to interpret European homelessness. The increased emphasis on individual characteristics and actions in theoretical debates about homelessness is considered and the importance of new arguments that homeless people exercise controls over the nature of their homelessness is explored. Finally, the paper discusses how to take forward analysis of a social problem that has both individual and structural elements.

>> Keywords_ Old and new orthodoxy, theory and data, theory building on homelessness in the EU

ISSN 2030-2762 / ISSN 2030-3106 online 20 European Journal of Homelessness _ Volume 10, No. 3 _ 2016

Introduction: The New Orthodoxy

Victorian investigations of extreme, unaccommodated poverty looked for individual moral lapses, active choices not to work honestly, to drink too much and to be criminal; or for individual tragedies, decent people finding themselves in desperate situations despite their very best efforts (Ribton-Turner, 1887; Freeman and Nelson, 2008; Higginbotham, 2008; O’Sullivan, 2016). Twentieth century scholars used statistics and ethnographic techniques to study homeless populations, defined as people using emergency accommodation and living rough. Data from this research repeatedly showed a very high prevalence of mental health problems, initially in combination with problematic drinking and, later on, with illegal drug use, within largely lone adult male populations (Scott, 1993). However, research into homeless families, who tended to be very poor but not seriously mentally ill; data that raised questions about how high the prevalence of support needs was among homeless individuals; and apparent spikes in homelessness linked to economic recession, downward shifts in affordable housing supply and cuts to welfare systems, caused some academics to redirect their attention to structural factors (Burt, 1991). Academic debates about homelessness can be characterised by arguments about whether structural factors, or individual pathology, provides the better explanation of why homelessness occurs (Gowan, 2010; Farrugia and Gerrard, 2015).

A ‘new orthodoxy’ arose in the 1990s and began to shape theoretical debates and the conduct of homelessness research (Caton, 1990; Pleace, 1998; 2000; Farrugia and Gerrard, 2015). This new orthodoxy contended that homelessness was not individual in nature, nor was it structural, but instead resulted from the interaction of structural and individual factors (O’Flaherty, 2004; Fitzpatrick, 2005). Structural factors, in the new orthodoxy, referred to the operation of housing and labour markets, cultural and political factors contributing to inequity – i.e., sexism, racism and other forms of stigmatisation; and to the operation of welfare, public health and social housing systems. Individual factors centred on needs, characteristics, experiences and, importantly, behaviour.

The new orthodoxy was posited on the idea that most homelessness research tended to suffer from one of two main theoretical flaws (Neale, 1997). The first theoretical flaw, which dated back to the first attempts at studying homeless people, was a near-total emphasis on observable individual traits (O’Sullivan, 2016). Homelessness was explained using analysis of specific people in specific places, which by and large meant people living rough or in emergency shelters, whose apparently very high support needs were used to ‘explain’ their homelessness, with little or no reference to contextual variables (Hopper, 1990; O’Sullivan, 2008). The second theoretical flaw in existing homelessness research – according to the new orthodoxy – was a second set of ideas that viewed homelessness as the polar 10th Anniversary Issue 21

opposite, as almost entirely a consequence of capitalism, in a context of failures within, and cuts to, welfare systems. Homelessness was seen as being inflicted on powerless people by forces that were, literally, beyond their control, or as a function of individual pathology (Pleace, 2000).

Within the new orthodoxy, labour markets, housing markets, welfare systems, health and social housing systems, and individual needs, behaviour, characteristics and experiences all combined to cause homelessness (Caton, 1990; Pleace, 2000; O’Flaherty, 2004). When someone experienced the wrong combination of structural and personal factors, homelessness was created and sustained. Homelessness, was a negative assemblage of structural and individual disadvantages; homelessness was a pattern (Lee et al., 2010).

In the new orthodoxy, three factors worked in combination. These were personal capacity, access to informal support and access to formal support.

The risk of homelessness increased if someone lacked personal capacity, which meant resilience, coping skills and access to financial resources. This meant labour market disadvantage, limiting illness, disability, low educational attainment, a disrupted childhood, mental illness, drug addiction, criminality – indeed, anything that limited someone’s innate capacity to self-care in a free market economy.

Limits to personal capacity could be countered by a partner, family or friends if they were able – and were prepared – to offer informal support. If someone could not put a roof over their own head, a family, a partner or a friend might do so. If there was no partner, friends or family, these supports were absent, as was the case if these potential sources of informal support had no resources to spare or were simply unwilling to help.

Formal support from health, welfare, social housing systems and homelessness services could, in turn, counteract limits to both personal capacity and to informal support. Barriers to health, care, support and housing services for homeless people, ranging from negative responses based on stereotypical images of homelessness through to local connection rules, could, however, block access to formal support (Baptista et al., 2015). Formal support services might also simply be under- resourced, which might in itself generate homelessness. This raised the possibility that countries with well resourced, highly accessible welfare systems would have less homelessness, an argument for which there is some evidence (Stephens and Fitzpatrick, 2007; Benjaminsen and Andrade, 2015).

Systemic failure – focused specifically on homelessness services rather than welfare systems as a whole – had also been identified by researchers as a cause of homelessness. It had been apparent for decades that some members of the homeless population were recurrent and long-term users of homelessness services 22 European Journal of Homelessness _ Volume 10, No. 3 _ 2016

(Macgregor-Wood, 1976; Dant and Deacon, 1989). However, the true nature of that population and how small it might actually be would not start to become clear until the ground-breaking longitudinal analysis of Culhane and his colleagues on service administrative data in the US (Culhane et al., 1994). These studies indicated that 10 percent of Americans using emergency shelters were long-term homeless people with high support needs, with another 10 percent characterised by recurrent homelessness and relatively high support needs (Kuhn and Culhane, 1998). Some people were not being brought out of homelessness by existing services, something that was found in EU Member States as well as in the USA (Sahlin, 2005).

A triad of support systems kept homelessness at bay: personal capacity, informal support and formal support. If one set of supports failed, homelessness might be avoided; remove two and the risk of homelessness increased; once all three were gone, homelessness was, from a new orthodoxy perspective, practically inevitable.

Structural factors were apparently at the core of the new orthodoxy: economic systems actively generated inequity; barriers to welfare, health and housing services and inadequate homelessness services were all contributory factors. Yet the factors that predicted – indeed, determined – homelessness were individual. Individual capacity, if it were sufficient, meant the risk of homelessness could be effectively resisted, particularly if combined with sufficient informal support from friends, family and a partner. Vulnerability to homelessness due to structural factors began with individual characteristics; severe mental illness, drug addiction, criminality, sustained worklessness and limiting illnesses all undermined individual capacity and they might also undermine or remove access to informal supports (Dant and Deacon, 1989). The new orthodoxy was posited on the idea that structural factors ‘caused’ homelessness only when someone had limits to their personal capacity and insufficient access to informal support. One had to need formal support to prevent or exit homelessness before inadequacies within, or barriers to, that formal support started to matter as a cause of homelessness (Pleace, 2000; Lee et al., 2010).

The new orthodoxy has influenced both research and practice. Academic research has explicitly linked homelessness service failures to a false construct of homelessness as the result of individual pathology, without sufficient acknowledgement of structural factors or a lack of informal support as causes of homelessness. Services have been criticised because they seek to ‘correct’ deviant behaviours, treat mental health problems and drug-addiction, and ‘staircase’ homeless people towards independent living through training and treatment (Pleace, 2008). Some American research found evidence of reprehensible practice and failure in staircase services (Stark, 1994; Lyon-Callo, 2000; Dordick, 2002), echoed in European analysis (Hutson and Liddiard, 1994; Sahlin, 2005; Busch-Geertsema and Sahlin, 2007). A more complex picture was suggested by other research, which reported 10th Anniversary Issue 23

‘staircase’ models that were supportive, choice-orientated and relatively effective (Pleace, 2008; Rosenheck, 2010), rather than total institutions (Stark, 1994). However, the idea that homelessness services failed because they greatly over- emphasised individual behaviour and had fixed, preconceived ideas about who homeless people were, became pervasive (Busch-Geertsema and Sahlin, 2007).

Equally, service responses that failed to acknowledge individual characteristics, needs, experiences and behaviour have also been subject to academic criticism. The ‘housing-only’ response of British statutory services, which offered secure social housing to vulnerable homeless individuals without support services has been criticised by researchers and practitioners as being effectively over-focused on housing need (Dant and Deacon, 1989; Pleace, 1995; McNaughton-Nicolls, 2009).

New service interventions that emphasise shared humanity and a respect for the choices of homeless people have appeared, the leading one of which is, of course, Housing First. In a Housing First service, homeless people are not blamed for their situation; their housing need is recognised as related to structural factors and is met, but there is at least equal emphasis on meeting individual support needs within a framework characterised by service user choice, harm reduction and a recovery orientation. Housing First is, in many senses, almost akin to a manifesta- tion of the new orthodoxy (Tsemberis, 1999; Hansen-Löfstrand and Juhila, 2012; Padgett et al., 2016; Pleace, 2016). By contrast, a traditional, basic emergency shelter can be directly related to the ideas within the old orthodoxies that homeless people mean less, have less and are less than ordinary citizens, which is ‘why’ they are homeless (Lancione, 2016).

Subgroups and Agency

The problem with the new orthodoxy is an absence of precision. If homelessness is, indeed, the result of a negative assemblage of individual and structural factors, questions then arise as to how exactly this happens and what it looks like. Neale’s criticism of arguments emphasising structural factors or individual factors is that homelessness is too diverse to support either set of assumptions. As she notes, this is an equally effective critique of the dangers of the imprecise melding of individual and structural factors within the new orthodoxy (Neale, 1997). Hopper also highlights the inherent vagueness of the new orthodoxy, arguing that both a single pattern and a set of patterns are hard to see within an essentially heterogeneous population (Hopper, 2003). Writing in 2000, the author similarly criticised the new orthodoxy for lacking a clear expository framework (Pleace, 2000). 24 European Journal of Homelessness _ Volume 10, No. 3 _ 2016

For critics, the new orthodoxy was not a testable hypothesis; it failed, even in broad terms, to explain how this causal interaction of personal and structural worked. The criticism was that the new orthodoxy amounted to a series of vague suggestions, not a coherent, testable, social scientific theory (Neale, 1997; Pleace, 2000; Williams and Cheal, 2001; Pleace and Quilgars, 2003; Somerville, 2013).

For Fitzpatrick, the way to move beyond these limits was to look at the detail, to zoom in on homelessness causation and explore whether distinct patterns were present. There might well not be one ‘homelessness’, triggered by a consistent negative assemblage of trigger variables, but there could be identifiable, predictable clusters of homelessness, taking differing forms. If ‘internally homogeneous subgroups’ existed, this allowed for the possibility that the individual factors would sometimes be more important than structural factors, or indeed vice versa, depending on which form of homelessness one was talking about (Fitzpatrick, 2005). According to these ideas, economic structures, housing structures, patriarchal and interpersonal structures and individual attributes do not manifest homelessness in one – ill-defined and unexplained – form, but as a series of distinct social problems (Williams and Cheal, 2001; Fitzpatrick, 2005). Subsequent work on ‘multiple-exclusion’ homelessness used statistical analysis to explore clusters of characteristics in a specific, long-term and recurrently homeless population, arguing that it did, indeed, have distinct, predictable characteristics (Fitzpatrick et al., 2011).

While Fitzpatrick and others have sought to refine the new orthodoxy, other voices have questioned it on a fundamental level. These criticisms centre on the idea that structural factors were being over-emphasised and that agency was effectively being removed from homeless people (McNaughton-Nicolls, 2009).

From this perspective, the new orthodoxy disempowers homeless people, both as agents whose decisions might negatively influence their situation and as agents whose individual actions might enable them to find their own way out of homelessness. The assumptions of old orthodoxies – that homeless people were powerless in the face of structural forces or could not overcome their own limitations due to issues like mental health problems – were still present.

From this perspective, homelessness is navigation; it is a navigation that may be constrained in various ways, but homeless people nevertheless take decisions that influence their trajectories through homelessness. Understanding that individual choices influence homelessness along with individual characteristics, needs and experiences as well as structural factors, leads to the idea that people take specific pathways through homelessness (Snow et al., 1994; May, 2000; Clapham, 2003; Fopp, 2009; McNaughton-Nicholls, 2009; Parsell and Parsell, 2012; Somerville, 2013). This line of criticism of the new orthodoxy is distinctive because the study of an individual homeless person is not simply a ‘diagnosis’ to explain their homeless- 10th Anniversary Issue 25

ness (Lyon-Callo, 2000; O’Sullivan, 2008). There is recognition of the homeless person as an agent – a thinking person taking a particular ‘pathway’ through homelessness (Clapham, 2003) and not simply someone who is a victim of their own vulnerabilities in a harsh world – while also acknowledging that structural factors are present. This emphasis on actions and behaviour brings individual characteristics to the fore, making understanding homelessness a matter of understanding individual choices to a much greater extent than is suggested by the new orthodoxy (McNaughton-Nicolls, 2009; Parsell and Parsell, 2012).

Theory and Evidence

Relatively primitive individualist explanations of homelessness, which over time shifted towards ethnography, started to be challenged by academics arguing that structural factors were fundamental to understanding homelessness. In turn, these two underdeveloped ideas were replaced by the new orthodoxy, in which negative assemblages of structural and individual factors were seen as creating homelessness, which was, in turn, criticised for lacking precision (Neale, 1997; Clapham, 2003; Fitzpatrick, 2005). The study of subgroups, in which homelessness stopped being one social problem and became many, combined with a recognition of individual agency, has begun the process of working towards an analytical framework for European homelessness research (Pleace, 2005).

However, homelessness research, both in the European context but also more broadly, is still uncomfortably close to being a conceptually inconsistent mess. Research focused on structural factors, on individual pathology, using the new orthodoxy, using a choice-focused ethnographic ‘pathways’ analysis and using subgroup analysis is occurring, and being published, at the time of writing.

Quite a lot of the research emphasising structural factors, such as inadequate housing supply as a direct cause of homelessness in the UK (Greve, 1964; Glastonbury, 1971; Greve et al., 1971; Drake, 1989) or highlighting rising inequality in the USA (McCarthy and Hagan, 1991; Burt, 1991), is older material. However, papers arguing that homelessness is essentially a construct of the neo-liberalist pursuit of inequity are still appearing (Phelan and Norris, 2008; Willse, 2010; Bullen, 2015). Ethnographic studies emphasising individual pathology are still published (Nooe and Patterson, 2010), as are papers that feel there is still a need, more than 25 years since the new orthodoxy first started to appear, to argue that homelessness has both individual and structural causes (Cronley, 2010; Lee et al., 2010; Parsell and Marston, 2012; Piat et al., 2014). 26 European Journal of Homelessness _ Volume 10, No. 3 _ 2016

Not all of the current work on homelessness acknowledges that the evidence base has undergone radical change in the last 25 years. Longitudinal analysis of large scale administrative datasets has shown there are patterns in American homelessness (Culhane et al., 1994; Kuhn and Culhane, 1998; Culhane et al., 2013). There is a small population with poor mental and physical health, and with limited or no informal support, who are not getting access to the right services. These studies found that cross-sectional research had over-sampled a high-need minority who were the most likely to be sleeping rough or in services, while anyone experiencing homelessness for a shorter period tended to be missed (O’Sullivan, 2008). Longitudinal analysis found a much larger, transitionally homeless population who were poor and who tended not to have high support needs (Culhane et al., 2013).

Evidence has since appeared indicating that clusters of high-need long-term homeless people exist elsewhere: in London (Jones and Pleace, 2010); Toronto (Aubry et al., 2013); Dublin (O’Donoghue-Hynes, 2015); and, particularly, in Denmark (Benjaminsen and Andrade, 2015). The scale of transitionally homeless populations appears to be very much smaller in some European contexts where welfare systems are extensive than in America, while high-need subgroups are still present (Benjaminsen, 2016). The same work also partially supported the arguments of those who said homelessness was broadly related to inequity due to evidence of a transitional, apparently low-need homeless population characterised primarily by poverty (Pleace, 1998; Farrugia and Gerrard, 2015).

Analysis has also showed that support needs and sets of behaviours associated with long-term homelessness do not always predate homelessness but can arise during homelessness. People who do not have high support needs, or indeed any support needs, when they first become homeless, develop support needs in association with experiencing sustained or recurrent homelessness. Peaks in homelessness also appear to be related to economic recessions, and American research has found that long-term and recurrently homeless people tend to be a similar age (Culhane et al., 2013). This work suggests that the flow into long-term and recurrent homelessness may not be constant; it could peak during recessions, which, when coupled with data indicating that high support needs can arise during homelessness, raises an interesting possibility: if these data are right, they mean people without high support needs become homeless at higher rates during recessions and that some of this group find themselves unable to exit homelessness and experience marked deteriorations in mental and physical health in association with what becomes long-term and repeat homelessness. 10th Anniversary Issue 27

Structural factors have become more evident through comparative analysis. Recent research from Denmark, contrasting results with the USA, indicates that the nature of homelessness can be distinct in countries with radically different welfare systems. Homelessness that is related simply to poverty appears to be very unusual in the Danish context, in marked contrast with the USA (Benjaminsen, 2016).

Criticism of the idea that severe mental illness was actually a demonstrable cause of homelessness – as people with severe mental illness greatly outnumber homeless people and mental illness can arise during homelessness – has been around for decades (Cohen and Thompson, 1992). Systemic failures in mental health treatment and welfare systems can be associated with homelessness, but current evidence indicates that poor mental health in itself is not a sufficient or necessary cause of homelessness (Montgomery et al., 2013). Research has also highlighted how drug use – sometimes described as another ‘causal’ factor – can arise during homelessness or remain constant, beginning before homelessness, continuing during homelessness and persisting after homelessness (Kemp et al., 2006; Johnson and Chamberlain, 2008; Pleace, 2008).

Domestic violence and abuse has also been found to be a far more frequent cause of women’s homelessness than is the case for male homelessness, but gender differences in homelessness causation appear to be only part of a larger picture. Women appear often to take distinct pathways through homelessness, which are not explained by differences in structural factors such as differences in welfare systems, but which are instead linked to agency. Families, disproportionately headed by lone women, appear often to respond to homelessness by relying on informal support, only resorting to services once support from friends or relatives becomes exhausted, according to American and British research (Shinn et al., 1998, Pleace et al., 2008). Women’s experience of homelessness when they were without or separated from children has been found to be similar, with reliance on personal and informal resources again leading to an experience of homelessness that often remains hidden, and with services again being used as a last, rather than a first, resort (Mayock and Sheridan, 2012; Mayock and Bretherton, 2016). Studies of youth homelessness have also shown how young people’s experience of homelessness can be shaped by how they respond to their situation. Again, they might use their own capacity and informal resources rather than going straight to services, at least when homelessness initially occurs (Fitzpatrick, 2000; Quilgars et al., 2008).

So, simple poverty can cause homelessness, and the extent to which this occurs can be linked to structural differences, such as those between welfare systems. Individual support needs, individual actions and levels of informal support may cause homelessness, allow homelessness to be avoided or result in varying trajectories through homelessness. When homelessness is recurrent or sustained, it can 28 European Journal of Homelessness _ Volume 10, No. 3 _ 2016

be associated with the emergence of a drug problem, mental illness or other support needs. There is no single route through homelessness, but apparent clusters exist, such as women experiencing homelessness trajectories that can be linked to their choices. Structural factors, individual agency, needs, characteristics and experiences have all been shown to have an influence on homelessness causation and the sustainment or recurrence of homelessness. This could be read as an argument that the simplicities of the new orthodoxy and the old orthodoxies look set to be replaced with a kind of ‘complex subgroups’ thesis (Pleace, 2005), but there are reasons to hesitate before going down this road.

The Limits of Current Theory

A key limitation in current theory is the extent and validity of observation. American research has delivered solid critical analysis of the limits in only looking at homeless people in specific contexts and at specific times (O’Sullivan, 2008; 2016). Yet, while the longitudinal research using administrative data conducted by Culhane and others in the US has been nothing short of ground-breaking, these are data based on service contacts; they are not the whole homeless population. Recent American work, involving Culhane, has begun to explore the possible extent of homeless populations beyond those who make contact with services, raising questions about the original thesis on the nature of American homelessness (Metraux et al., 2016).

The literature produced in the US is gigantic, but alongside a mix of robust, careful quantitative analysis and carefully conducted and contextualised ethnography, there are a lot of programme and service evaluations. A lot of US research is centred on exploring how to reduce homelessness by testing different models, not exploring the nature of homelessness itself. Canadian and Australian research has also added to scholarly discussion on the nature of homelessness (Parsell and Parsell, 2012; Piat et al., 2014), although, as in the US, quite of lot of the research being conducted is centred on programme and service model evaluation.

Good quality American, Australian and Canadian research adds to our understanding. However, this material is ultimately not about Europe, and that, in itself, is an important caveat. There is a need for caution in relying on externally generated evidence and ideas to guide European research, because it is already clear that patterns in homelessness that exist in countries outside Europe do not necessarily exist in the same way within Europe (Benjaminsen and Andrade, 2015).

There has been progress in developing an evidence base in Europe. France, Spain and Italy have all undertaken significant, if only periodic, attempts to count their homeless populations. Data on homelessness have also improved in countries like Poland and Portugal (Busch-Geertsema et al., 2014). In the United Kingdom, home- 10th Anniversary Issue 29

lessness research is extensive, although predominately funded by governments and the charitable sector, both of which are pursuing specific agendas and, again, heavily focused on service evaluation (Pleace and Quilgars, 2003). Denmark and Ireland have integrated data systems that, while not providing every answer, enhance statistical data on homelessness.

Nevertheless, European data on homelessness are skewed. There are geographical gaps. There are gaps in evidence on various forms of homelessness. The evidence base still tends towards studies of visible homelessness – i.e., populations largely comprised of lone men living rough and in homelessness services (Busch-Geertsema et al., 2010; Busch-Geertsema et al., 2014). People living without their own space, without privacy and without security of tenure in Europe are, at best, partially mapped and partially understood. European data on hidden or concealed homelessness, which includes youth, family and women’s homelessness, are quite limited.

Gaps in data mean gaps in understanding. Not knowing about a population makes it difficult to theorize about that population and creates the risk – as was the case in the US – of building theories about homelessness that simply fall apart as soon as data improve (O’Sullivan, 2008).

Issues around observation exist alongside problems with definition in the European context. Definitional challenges for researchers and policy-makers exist at two levels. The first is when homelessness is reduced to an ill-evidenced, over-simplified construct. For example, assuming that ‘all homeless people are mentally ill’ means that research, strategy and services are, at best, only covering one group and are, at worst, wasting resources and causing distress as they attempt to understand and respond to homelessness solely in terms of ‘mental health’ (Pleace, 2005). Equally, simply viewing homelessness as a function of housing market failure, corrected by increasing affordable housing supply, is also a flawed response, as assuming homeless people have no support needs is no better than assuming they are defined solely in terms of support needs. At the second level is the question of what is meant by homelessness itself, questioning why someone on the street or in an emergency shelter is ‘homeless’, while someone squatting in a building unfit for habitation or living in a shanty town is not. Concealed homeless households, without the privacy, safety or security of tenure that would be in place if they had their own front door, are ‘homeless’ in one European country but are only ‘badly housed’ in another (Busch-Geertsema et al., 2014).

The work of FEANTSA and the European Observatory on Homelessness in the MPHASIS project and developing the ETHOS typology has promoted the idea of a shared European definition of homelessness. Some progress has been made towards a universal standard for enumeration (Busch-Geertsema et al., 2014). Yet, 30 European Journal of Homelessness _ Volume 10, No. 3 _ 2016

beyond people living rough and, usually, in emergency accommodation, definitions of homelessness can still be inconsistent and contested, both within Europe and elsewhere (Amore et al., 2011).

Definition has major implications for theory. As the definition of homelessness broadens, structural factors may become more prominent, because more and more poor people enter the equation. It is the recurrently and long-term homeless populations that, on current evidence, have consistently high support needs, whereas other homeless people, such as families or those experiencing short term homelessness, do not (Burt, 1991; Kuhn and Culhane, 1998; Pleace et al., 2008).

Counting, say, concealed or hidden households as homeless, alongside people living rough, extends homelessness, even in contexts in which welfare and social housing systems are extensive and well-resourced (Benjaminsen, 2016). Homeless families do not look like people who have been living rough for any amount of time (Pleace et al., 2008). A precariously housed group, or indeed groups, that transition in and out of homelessness for short periods, if they are counted as homeless, will be different again and may in turn contain more subgroups (Culhane et al., 1994; Meert and Bourgeois, 2005). The further one moves away from regarding homelessness as only meaning long-term and recurrently homeless people, the more complex the picture, potentially, becomes. As the definition of the homeless population expands, new characteristics, new sets of behaviours and new structural factors will be added to the mix, and one theory may, in consequence, need to give way to another (Pleace, 2005).

Analysing homelessness as subgroups or as sets of pathways provides one way to try to tackle this issue, as it breaks homelessness up into more manageable conceptual chunks. However, taxonomies always have some element of compro- mise; there are ‘boundary’ cases that could go into one category or another, and decisions about the criteria used to identify each subgroup and whether it represents a robust basis for analysis are rarely straightforward (Bowker and Leigh-Star, 2000). Building a taxonomy becomes more and more complex the broader the definition of a social problem is and the more extensive the data are. Building clear and consistent pathways or subgroups is likely to be difficult in a data-rich environment with a wide definition of homelessness. Recent work from the US has shown how adding new data can disrupt taxonomies that were assumed to be relatively robust (Metraux et al., 2016).

Taking something like lone adult homelessness, it can be seen how one presumed pathway – from psychiatric ward to homelessness service – was fractured as data improved. The idea that drug and alcohol use, combined with mental health problems, prompted most lone adult homelessness also fell apart once it was seen that these issues could arise following homelessness and that many homeless 10th Anniversary Issue 31

people did not have these characteristics. Once gender was examined, it was clear that individual agency – women avoiding homelessness services, modifying their own trajectories through homelessness – also has an impact. Comparative research on welfare regimes shows the importance of context.

The potential problem with subgroups is that, as more data are added and as definitions widen, existing assumptions and existing patterns may disappear, supposedly ‘shared’ characteristics being replaced by more complex and nuanced relationships. Enough complexity in data may cause a breakdown in existing taxonomies, which as Neale (1997) suggests, can collapse in the face of enough intricacy. The point is that, if instead of, say, ten subgroups that provide a conceptual framework for ten distinct homelessness populations, there are a thousand, at what level of ‘membership’ does a homeless subgroup cease to be of theoretical or practical use? Equally, if there are many similarities in homeless populations, classification becomes a problem because there is not enough diversity.

Europe and the Conceptual Life Raft

Limitations in data have led to a tendency among European academics to use American, Australian and Canadian data and theory as a kind of conceptual life raft. European theoretical work on homelessness does exist but, inevitably, this work draws on American ideas because that is where most of the research and, consequently, much of the thinking about homelessness is done. The new orthodoxy was being written about in the US 26 years ago (Caton, 1990), while the importance of understanding behavioural factors, of pathways through homelessness, was being discussed 22 years ago (Snow et al., 1994). European academic thought on homelessness is not plagiarism, but it would be disingenuous to suggest that European homelessness research is not heavily influenced by American work. While European data have improved, it is still the case that the only data on specific aspects of homelessness, on (what may be) specific subgroups of homeless people, or sometimes the only research that is socially scientifically robust, is American, Australian or Canadian (Busch-Geertsema et al., 2010).

In Denmark, data-merging has already allowed American ideas about homelessness to be tested, with some very important differences in the nature of homelessness becoming evident (Benjaminsen and Andrade, 2015). In Ireland, the Pathway Accommodation and Support System (PASS), introduced in 2013, also allows for this kind of analysis. Better data allows testing of American ideas, which may show their limits in the European context and spur the development of new theories. Building a European theoretical debate about the nature of homelessness, adding to global academic discourse on homelessness, means undertaking more research 32 European Journal of Homelessness _ Volume 10, No. 3 _ 2016

in some contexts while redirecting efforts in others, particularly with respect to shifting the focus of research away from men who are living rough or in emergency accommodation.

There are reasons why the European evidence base is not more extensive. Quite a lot of homelessness research is funded by charities or governments and is focused on evaluation, or seeking to highlight specific issues. Perhaps more importantly, social scientific attention and the resources available to social scientists are confronted with a great many social problems. In the context of massive, structural shifts in labour markets producing mass youth unemployment across Europe; managing a population that is now living very long lives; or managing the consequences of mass migration, homelessness can appear to be a relatively minor issue, despite the unique the level of distress it causes.

Redirection of current research efforts, which while not necessarily large in scale or as robust as would be ideal, might help lessen Europe’s collective reliance on internationally provided conceptual life rafts. Looking again at homelessness causation, testing the viability of current theory, and exploring – albeit sometimes in smaller-scale work – the lived experience, perceptions and experience of homeless people is one way forward. There are dangers in swimming to these life rafts because of an absence of data: first, American, Australian and Canadian ideas might not be relatable to European contexts; and second, there is a risk of projecting externally developed taxonomies and theories about the nature of homelessness onto European homeless people without sufficient critical analysis.

There is something to learn from those Americans who have conducted ethnographic research that has enriched and contextualised an understanding of homelessness, which is often built around statistical data. Working towards a mixed, multidisciplinary evidence base in Europe, including through ethnographic research, will help develop theoretical thinking, reduce the risk of incorrectly ‘projecting’ American patterns of homelessness onto a European context, and ensure homeless people and their views and experience are represented (see, for example, Lancione, 2016).

Large scale statistical data are vitally important; analysis of major administrative data sets, as achieved in Denmark (Benjaminsen, 2016), could revolutionize understanding of European homelessness. Yet, American experience teaches us to test the limits of administrative data and reflect on the lived experience of homelessness as well. The work of those who have emphasised individual agency in understanding European homelessness – for example from evidence about choice and differentiated experiences between genders (Mayock and Bretherton, 2016) – also highlights this need. 10th Anniversary Issue 33

Assumptive Research and Cultural Gravity

Early studies of homelessness took place in civilisations that, broadly, regarded themselves as natural systems, in which every individual had a place, usually at the bottom of a strict hierarchy. Taxonomies, including some spectacular categorisa- tions of ‘professional beggars’, ‘lunatics’ and ‘habitual drunkards’, were constructed to try to understand this population (Ribton-Turner, 1887; Gray, 1931).

Welfare systems were in place for populations we would now call homeless, using a mixture of containment and support (Roberts, 1963; Higginbotham, 2008; O’Sullivan, 2016). These systems were not welfare states, but the idea that poverty was, at least in part, structural in origin and that society had obligations towards the poor extended beyond the Marxists, even if the idea that some of the poor were ‘undeserving’ proved very hard to shake off (Vorspan, 1977; Veit-Wilson, 1991).

In the context of highly developed welfare systems that were built on assumptions that the poor and vulnerable needed to be cared for by the state – in Western Europe at least – a shift in homelessness research occurred. Research began to appear that conceptualised homelessness in terms of inadequate access to support, to housing, as a systems failure rather than as the result of individual action (Macgregor-Wood, 1976). In making the argument that homeless people were not being properly cared for, that their rights to housing or support were not being recognised, these ideas created a new kind of individual pathology centred on a need for support. The presumption that asylum closures had put people with mental health problems onto the street and ‘caused’ the increase in American homelessness was the meridian of this kind of thinking (Scott, 1993).

Arguments that homelessness resulted mainly from labour market (Stewart, 1975) and housing market failure (Glastonbury, 1971; Greve, 1964; Greve et al., 1971; Drake et al., 1981) were relatively unusual and relatively short-lived. In the UK, more or less entirely economic arguments about the nature of homelessness were sometimes made (Drake, 1989). However, explanations of homelessness that emphasised the role of structural factors but also noted the role of individual characteristics were more common (Anderson, 1993).

Within the new orthodoxy itself, the role of individual characteristics in homelessness causation is fundamental. With the advent of arguments in favour of subgroup and pathways analysis, alongside the study of individual agency, the focus on the individual in homelessness has, if anything, increased.

Reading some of this literature, it can seem that things have moved on; the ‘sin-talk’ of homelessness as a moral offence, the ‘sick-talk’ of homelessness as a pathology and the ‘system-talk’ of homelessness as systemic injustice have been replaced 34 European Journal of Homelessness _ Volume 10, No. 3 _ 2016

(Gowan, 2010). Yet, mainstream responses to homelessness are another matter; as both American (Gowan, 2010) and European researchers have pointed out, individual pathology remains prominent in policy circles (Phillips, 2000; Pleace, 2000).

Images of homelessness as a problem experienced by specific types of individuals whose actions led to their homelessness were firmly within the cultural mainstream far into the twentieth century (Phillips, 2000) and remain present in Europe and throughout the Western world (Fopp, 2009; Devereux, 2015). Research in the meantime has, in some cases, remained focused simply on individual pathology or, while acknowledging structural factors, used a pathways or subgroup thesis (in which patterns are defined with reference to individual characteristics and behaviour) as a conceptual framework.

There are two risks here. These can be described as ‘assumptive research’ and a ‘cultural gravity’ risk, which could lead some European homelessness research in the wrong direction.

Assumptive research occurs when researchers regard homelessness as a clearly defined and understood social problem – i.e., that it is people sleeping rough who are largely male and whose homelessness is linked to support needs and behavioural factors. Such research adds nothing to the understanding of homelessness because it assumes homelessness is understood, which means that questions about the nature of homelessness do not need to be resolved, beyond determining, for example, how many rough sleepers in a particular city are taking heroin. People living without their own space, without privacy and without security of tenure in Europe are not considered by such research, because homelessness means a predominantly male population sleeping rough or in emergency shelters and nothing else.

One of the biggest challenges faced by European homelessness researchers is ensuring that there is a theoretical debate. Homelessness research is being conducted that assumes homelessness is a relatively simple, relatively small-scale social problem with clearly understood causes. There are longstanding concerns that the political right has sought to narrow the definition of what homelessness is, emphasising only visible homelessness that can be easily linked to individual pathology and drawing attention away from wider structural problems with affordable housing supply and inequity (Anderson, 1993). Beyond this, there is the view of homelessness as individual pathology that dates from before the nineteenth century – a mass cultural understanding of homelessness encompassing only a self-inflicted state found among people in emergency shelters and on the street (Carlen, 1996; Phillips, 2000; Gowan, 2010). Assumptive research must be challenged because it is based on a clearly false construct of what homelessness is and lacks any social scientific foundation. 10th Anniversary Issue 35

The ‘cultural gravity’ issue centres on context and on measurement. Homelessness research that makes allowance for structural factors, also taking into account individual characteristics and behaviour, goes against a cultural norm that sees homelessness in terms of individual pathology. Data tend to be collected at the individual level: what has happened to someone, their characteristics, their earlier experiences, their choices and so forth. Their experience of using systems, their experience with landlords, their interaction with the world in which they are homeless, can be asked about but it is rarely observed directly. The systemic or structural can be harder to see, particularly in a European context where large-scale, robust longitudinal research on the nature of homelessness is unusual. Conducting research in a context in which individual pathology is expected to be the causal factor, in which data can only be collected at the individual level, usually using cross-sectional methods or in a single interview, may make structural factors inherently harder to see (Farrugia and Gerrard, 2015).

Exploring Structural Causation

During the decade for which this journal has existed, our understanding of homelessness has moved beyond individual pathology, structuralism and beyond the new orthodoxy. There are challenges around definition and data quality, and in trying to avoid reaching across the Atlantic for reassurance and guidance. Ensuring the individual and their agency is represented and ensuring that representation is accurate is vital. Part of that challenge, alongside ensuring homeless people have their own voice and do not have ideas projected onto them, is to look at how the context in which homelessness occurs influences the people who experience it.

So, is there, then, a case for a reassertion of exploring structural factors in homelessness causation, countering all this nasty historically and culturally generated individual pathology? Equally, is there not a case for reasserting the role of structural factors to reduce the risk that researchers using pathways and subgroup conceptual frameworks get carried away and focus too much on the individual and not enough on structural factors?

On the surface, the risks of this idea are the same as they have always been – i.e., too much emphasis on structural variables risks potential distortion through not paying sufficient attention to the role of individual agency or individual characteristics. However, as argued above, there are clear risks in over-emphasising individual agency and characteristics in trying to understand homelessness. These risks centre on how far a subgroup or pathways approach can go. If presented with too much complexity, as Neale pointed out twenty years ago (Neale, 1997), or if presented with too much similarity, taxonomic approaches to homelessness may 36 European Journal of Homelessness _ Volume 10, No. 3 _ 2016

fall over. Focusing too much on the individual can risk downplaying structural factors, but perhaps equally significantly, taxonomies might not work very well, creating temptations to avoid some data (everyone is poor) or defining only specific groups as ‘homeless’ (recurrently and long-term homeless people are differentiated by high rates of severe mental illness) to make a taxonomy work.

We are clearly not yet in the position, in terms of data richness or moving towards a wider definition of European homelessness, where an existential risk to pathways or subgroup based analysis exists. It may be the case that homelessness is less complex in nature than current research suggests it may be, despite growing evidence of multiple variables at multiple levels having roles in the nature of the experience.

Almost 20 years ago, this author argued that homelessness was not a discrete social problem, but was instead an extreme of poverty being misread as ‘unique’. Specific interventions and support services were needed, as homeless people had particular needs, characteristics and experiences, but homelessness was, ultimately, a product of poverty, which meant effective policy had to tackle deeply engrained inequality (Pleace, 1998). Revisiting that argument now, there are flaws: there was no consideration of individual agency; homelessness was effectively – as in the new orthodoxy – equated with total disempowerment, which was clearly wrong, while individual characteristics were downplayed. However, looking at that paper again, the challenge it posed – that homelessness needs to be contextualised to be fully understood – does not seem to have been fully answered (see also Farrugia and Gerrard, 2015).

Much of the development of social science, including what development there has been in the academic study of homelessness, took place in what looks increasingly like an historically exceptional period. In the last 30 years, the tendency towards massive accumulation of wealth and political power among small elites, which characterises most of human history, has reasserted itself. London, along with other European capitals and major cities, has become ‘Pikettyville’, in which a tiny transnational urban elite is shaping the nature of the city itself, creating enclaves of ‘Alpha territory’ (Burrows et al., 2016). This massive concentration of wealth has taken place alongside the constriction of full time, well-paid, secure work, sustained reductions in welfare spending and marked decreases in affordable housing supply.

Homelessness does appear to be linked to individual characteristics, to behaviour, to choices; it is not simply a matter of economics or housing supply. The people who experience homelessness have at least some control over what happens to them. Yet, can the ‘Alpha territories’ exist without it meaning something for poverty and, alongside poverty, for homelessness? As wealth and power become ever more concentrated and inequity increases within Europe, does that have an impact on 10th Anniversary Issue 37

the nature, extent and experience of homelessness? There are other questions, too, around ethnic, cultural and gender inequalities, which are structural, and which may also influence the nature and experience of homelessness (Pleace, 2011; Mayock and Bretherton, 2016). These are not questions we can explore properly if we focus too heavily on attempts to create taxonomies of homelessness as our main method of understanding this most acute of social problems, or as a set of interconnected social problems.

Our challenge as researchers and as social scientists is to fully acknowledge, respect and understand the human beings at the heart of homelessness and to understand as much as possible about the environment in which homelessness occurs. This requires a new neutrality, an openness, leaving behind preconceptions and ideas and theories about what we think homelessness is, who we think homeless people are and how we situate homelessness within the wider social and economic context. 38 European Journal of Homelessness _ Volume 10, No. 3 _ 2016

>> References

Amore, K., Baker, M. and Howden-Chapman, P. (2011) The ETHOS Definition and Classification of Homelessness: An Analysis, European Journal of Homelessness 5(2) pp.19-37.

Anderson, I. (1993) Housing Policy and Street Homelessness in Britain, Housing Studies 8(1) pp.17-28.

Aubry, T., Farrell, S., Hwang, S.W. and Calhoun, M. (2013) Identifying the Patterns of Emergency Shelter Stays of Single Individuals in Canadian Cities of Different Sizes, Housing Studies 28(6) pp.910 -927.

Baptista, I., Benjaminsen, L. and Pleace, N. (2015) Local Connection: Rules and Access to Homelessness Services in Europe (Brussels: FEANTSA).

Benjaminsen, L. (2016) Homelessness in a Scandinavian Welfare State: The Risk of Shelter Use in the Danish Adult Population, Urban Studies 53(10) pp.2041-2063.

Benjaminsen, L. and Andrade, S.B. (2015) Testing a Typology of Homelessness Across Welfare Regimes: Shelter Use in Denmark and the USA, Housing Studies 30(6) pp.858-876.

Bowker, G.C. and Leigh-Star, S. (2000) Sorting Things Out: Classification and its Consequences (London: MIT Press).

Bullen, J. (2015) Governing Homelessness: The Discursive and Institutional Construction of Homelessness in Australia, Housing, Theory and Society 32(2) pp.218-239.

Burrows, R., Webber, R. and Atkinson, R. (2016) ‘Welcome to ‘Pikettyville’? Mapping London’s Alpha Territories, The Sociological Review.

Burt, M.R. (1991) Causes of the Growth of Homelessness During the 1980s, Housing Policy Debate 2(3) pp.901-936.

Busch-Geertsema, V. and Sahlin, I. (2007) The Role of Hostels and Temporary Accommodation, European Journal of Homelessness 1 pp.67-93.

Busch-Geertsema, V., Edgar, W., O’Sullivan, E. and Pleace, N. (2010) Homelessness and Homeless Policies in Europe: Lessons from Research (Brussels: Directorate-General for Employment, Social Affairs and Equal Opportunities). 10th Anniversary Issue 39

Busch-Geertsema, V., Benjaminsen, L., Filipovič Hrast, M. and Pleace, N. (2014) The Extent and Profile of Homelessness in European Member States: A Statistical Update (Brussels: FEANTSA).

Carlen, P. (1996) Jigsaw: A Political Criminology of Youth Homelessness (Buckingham: Open University Press).

Caton, C. (1990) Homeless in America (Oxford: Oxford University Press).

Clapham, D. (2003) Pathways Approaches to Homelessness Research, Journal of Community and Applied Social Psychology 13 pp.119-127.

Cohen, C.I. and Thompson, K.S. (1992) Homeless Mentally Ill or Mentally Ill Homeless? American Journal of Psychiatry 149(6) pp.816-823.

Cronley, C. (2010) Unravelling the Social Construction of Homelessness, Journal of Human Behaviour in the Social Environment 20(2) pp.319-333.

Culhane, D.P, Dejowski, E.F, Ibanez, J., Needham, E. and Macchia, I. (1994) Public Shelter Admission Rates in Philadelphia and New York City: The Implications of Turnover for Sheltered Population Counts, Housing Policy Debate 5(2) pp.107-40.

Culhane, D.P., Metraux, S., Byrne, T., Stino, M. and Bainbridge, J. (2013) The Age Structure of Contemporary Homelessness: Evidence and Implications for Public Policy, Analyses of Social Issues and Public Policy 13(1) pp.228-244.

Dant, T. and Deacon, A. (1989) Hostels to Homes (Aldershot: Avebury).

Devereux, E. (2015) Thinking Outside the Charity Box: Media Coverage of Homelessness, European Journal of Homelessness 9(2) pp.261-273.

Dordick, G.A. (2002) Recovering from Homelessness: Determining the “Quality of Sobriety” in a Transitional Housing Program, Qualitative Sociology 25(1) pp.7-32.

Drake, M. (1989) Fifteen Years of Homelessness in the UK, Housing Studies 4(2) pp.119 -127.

Drake, M., O’Brien, M. and Biebuyck, T. (1981) Single and Homeless (London: HMSO).

Farrugia, D. and Gerrard, J. (2015) Academic Knowledge and Contemporary Poverty: The Politics of Homelessness Research, Sociology 50(2) pp.267-284.

Fitzpatrick, S. (2000) Youth Homelessness (London: Palgrave Macmillan).

Fitzpatrick, S. (2005) Explaining Homelessness: A Critical Realist Perspective, Housing, Theory and Society 22(1) pp.1-17. 40 European Journal of Homelessness _ Volume 10, No. 3 _ 2016

Fitzpatrick, S., Johnsen, S. and White, M. (2011) Multiple Exclusion Homelessness in the UK: Key Patterns and Intersections, Social Policy and Society 10(4) pp.501-512.

Fopp, R. (2009) Metaphors in Homelessness Discourse and Research: Exploring “Pathways”, “Careers” and “Safety Nets”, Housing, Theory and Society 26(4) pp.271-291.

Freeman, M. and Nelson, G. (Eds.) (2008) Vicarious Vagrants: Incognito Social Explorers and the Homeless in England, 1860-1910 (Lambertville, NJ: The True Bill Press).

Glastonbury, B. (1971) Homeless Near a Thousand Homes (London: George Allen and Unwin).

Gowan, T. (2010) Hobos, Hustlers, and Backsliders: Homeless in San Francisco. (Minneapolis: University of Minnesota Press).

Gray, F. (1931) The Tramp: His Meaning and Being (London: Dent).

Greve, J. (1964) London’s Homeless – Occasional Papers on Social Administration, Number 10 (London: G. Bell & Sons).

Greve, J., Page, D. and Greve, S. (1971) Homelessness in London (Edinburgh: Scottish Academic Press).

Hansen Lofstrand, C. and Juhila, K. (2012) The Discourse of Consumer Choice in the Pathways Housing First Model, European Journal of Homelessness 6(2) pp.47-68.

Higginbotham, P. (2008) The Workhouse Cookbook (Stroud: The History Press).

Hopper, K. (1990) Homeless Men in Historical Perspective, Journal of Social Issues 46(4) pp.13-29.

Hopper, K. (2003) Reckoning with Homelessness (Ithaca: Cornell University Press).

Hutson, S. and Liddiard, M. (1994) Youth Homelessness: The Construction of a Social Issue (Basingstoke: Macmillan).

Kemp, P.A., Neale, J. and Robertson, M. (2006) Homelessness Among Problem Drug Users: Prevalence, Risk Factors and Trigger Events, Health & Social Care in the Community 14(4) pp.319-28. 10th Anniversary Issue 41

Kuhn, R. and Culhane, D.P. (1998) Applying Cluster Analysis to Test a Typology of Homelessness by Pattern of Shelter Utilization: Results from the Analysis of Administrative Data, American Journal of Community Psychology 26(2) pp.207-232.

Johnson, G. and Chamberlain, C. (2008) Homelessness and Substance Abuse: Which Comes First?, Australian Social Work 61(4) pp.342-356.

Jones, A. and Pleace, N. (2010) A Review of Single Homelessness in the UK 2000 – 2010 (London: Crisis).

Lancione, M. (2016) Racialised Dissatisfaction: Homelessness Management and the Everyday Assemblage of Difference, Transactions of the Institute of British Geographers 41 pp.363-375.

Lee, B.A., Tyler, K.A. and Wright, J.D. (2010) The New Homelessness Revisited, Annual Review of Sociology 36 pp.501-521.

Lyon-Callo, V. (2000) Medicalizing Homelessness: The Production of Self-blame and Self-governing within Homeless Shelters, Medical Anthropology Quarterly 14(3) pp.328-345.

Macgregor-Wood, S. (1976) Camberwell Reception Centre: A Consideration of the Need for Health and Social Services of Homeless Single Men, Journal of Social Policy 5(4) pp.389-399.

May, J. (2000) Housing Histories and Homeless Careers: A Biographical Approach, Housing Studies 15(4) pp.613-638.

Mayock, P. and Bretherton, J. (Eds.) (2016) Women’s Homelessness in Europe (London: Palgrave Macmillan).

Mayock, P. and Sheridan, S. (2012) Women’s ‘Journeys’ to Homelessness: Key Findings from a Biographical Study of Homeless Women in Ireland. Women and Homelessness in Ireland, Research Paper 1. (Dublin: School of Social Work and Social Policy and Children’s Research Centre, Trinity College Dublin).

McCarthy, B. and Hagan, J. (1991) Homelessness: A Criminogenic Situation? British Journal of Criminology 31(4) pp.393-410.

McNaughton-Nicholls, C.M. (2009) Agency, Transgression and the Causation of Homelessness: A Contextualised Rational Action Analysis, European Journal of Housing Policy 9(1) pp.69-84.

Meert, H. and Bourgeois, M. (2005) Between Rural and Urban Slums: A Geography of Pathways through Homelessness, Housing Studies 20(1) pp.107-125. 42 European Journal of Homelessness _ Volume 10, No. 3 _ 2016

Metraux, S., Manjelievskaia, J., Treglia, D., Hoffman, R., Culhane, D.P. and Ku, B.S. (2016) Posthumously Assessing a Homeless Population: Services Use and Characteristics, Psychiatric Services 67(12) pp.1334-1339.

Montgomery, A.E., Metraux, S. and Culhane, D.P. (2013) Rethinking Homelessness Prevention among Persons with Serious Mental Illness, Social Issues and Policy Review 7(1) pp.58-82.

Neale, J. (1997) Homelessness and Theory Reconsidered, Housing Studies 12(1) pp.47-61.

Nooe, R.M. and Patterson, D.A. (2010) The Ecology of Homelessness, Journal of Human Behaviour in the Social Environment 20(2) pp.105-152.

O’Donoghue-Hynes, B. (2015) Patterns of homeless emergency accommodation use in Dublin: how do we compare? Presentation given at Families, Housing and Homelessness Dublin: Tenth European Research Conference on Homelessness 25th September 2015 http://www.feantsaresearch.org/IMG/pdf/ws_8_odonough_ patternsofhomelessea.pdf (accessed 10th November, 2016).

O’Flaherty, B. (2004) Wrong Person and Wrong Place: For Homelessness, the Conjunction is what Matters, Journal of Housing Economics 13(1) pp.1-15.

O’Sullivan, E. (2008) Pathways Through Homelessness: Theoretical and Policy Implications, in: J. Doherty and B. Edgar (Eds.) ‘In My Caravan, I Feel Like Superman’: Essays in Honour of Henk Meert, 1963-2006, pp.79–108. (Brussels: FEANTSA/Centre for Housing Research, University of St. Andrews).

O’Sullivan, E. (2016) Women’s Homelessness: An Historical Perspective, in: P. Mayock and J. Bretherton (Eds.) Women’s Homelessness in Europe (London: Palgrave Macmillan).

Padgett, D., Henwood, B. and Tsemberis, S. (2016) Housing First: Ending Homelessness, Transforming Systems, and Changing Lives (Oxford: New York).

Parsell, C. and Marston, G. (2012) Beyond the ‘At Risk’ Individual: Housing and the Eradication of Poverty to Prevent Homelessness, Australian Journal of Public Administration 71(1) pp.33-44.

Parsell, C. and Parsell, M. (2012) Homelessness as a Choice, Housing, Theory and Society 29(4) pp.420-434.

Phelan, E. and Norris, M. (2008) Neo-corporatist Governance of Homelessness Services in Dublin: Reconceptualization, Incorporation and Exclusion, Critical Social Policy 28(1) pp.51-73. 10th Anniversary Issue 43

Phillips, R. (2000) Politics of Reading: Cultural Politics of Homelessness, Antipode 32(4) pp.429-462.

Piat, M., Polvere, L., Kirst, M., Voronka, J., Zabkiewicz, D., Plante, M.C., Isaak, C., Nolin, D., Nelson, G. and Goering, P. (2014) Pathways into Homelessness: Understanding how both Individual and Structural Factors Contribute to and Sustain Homelessness in Canada, Urban Studies 52(13) pp.2366-2382.

Pleace, N. (1995) Housing Vulnerable Single Homeless People (York: University of York).

Pleace, N. (1998) Single Homelessness as Social Exclusion: The Unique and the Extreme, Social Policy and Administration 32(1) pp.46-59.

Pleace, N. (2000) The New Consensus, the Old Consensus and the Provision of Services for People Sleeping Rough, Housing Studies 15(4) pp.581-594.

Pleace, N. (2005) State, Trait, Choice or Something Else? The Need for a New Definition of Homelessness in the UK, Housing Studies Association Conference, Workshop Paper, York.

Pleace, N. (2008) Effective Services for Substance Misuse and Homelessness in Scotland: Evidence from an International Review (Edinburgh: Scottish Government).

Pleace, N. (2011) Immigration and Homelessness, in: E. O’Sullivan, V. Busch- Geertsema, D. Quilgars and N. Pleace (Eds.) Homelessness Research in Europe, pp.143-163. (Brussels: FEANTSA).

Pleace, N. (2016) Housing First Guide Europe (Brussels: FEANTSA).

Pleace, N. and Quilgars, D. (2003) Led Rather than Leading? Research on Homelessness in Britain, Journal of Community & Applied Social Psychology 13(2) pp.187-196.

Pleace, N., Fitzpatrick, S., Johnsen, S., Quilgars, D. and Sanderson, D. (2008) Statutory Homelessness in England: The Experience of Families and 16-17 Year Olds (London: Department for Communities and Local Government).

Quilgars, D., Johnsen, S. and Pleace, N. (2008) Review of Youth Homelessness in the UK (York: Joseph Rowntree Foundation).

Ribton-Turner, C.J. (1887) A History of Vagrants and Vagrancy and Beggars and Begging (London: Chapman and Hall).

Roberts, D. (1963) How Cruel was the Victorian Poor Law?, The Historical Journal 6(1) pp.97-107. 44 European Journal of Homelessness _ Volume 10, No. 3 _ 2016

Rosenheck, R. (2010) Service Models and Mental Health Problems: Cost Effectiveness and Policy Relevance, in: I.G. Ellen and B. O’Flaherty (Eds.) How to House the Homeless, pp.17-36. (New York: Russell Sage Foundation).

Sahlin, I. (2005) The Staircase of Transition: Survival Through Failure, Innovation 18(2) pp.115-136.

Scott, J. (1993) Homelessness and Mental Illness, British Journal of Psychiatry 162 pp.314-324.

Shinn, M., Weitzman, B.C., Stojanovic, D., Knickman, J.R., Jimenez, L., Duchon, L., James, S. and Krantz, D.H. (1998) Predictors of Homelessness among Families in New York City: From Shelter Request to Housing Stability, American Journal of Public Health 88(11) pp.1651-1657.

Snow, D.A., Anderson, L. and Koegel, P. (1994) Distorting Tendencies in Research on the Homeless, American Behavioural Scientist 37(4) pp.461-475.

Somerville, P. (2013) Understanding Homelessness, Housing, Theory and Society 30(4) pp.384-415.

Stark, L.R. (1994) The Shelter as “Total Institution”:” An Organizational Barrier to Remedying Homelessness, American Behavioural Scientist 37(4) pp.553-562.

Stephens M. and Fitzpatrick, S. (2007) Welfare Regimes, Housing Systems and Homelessness. How are they Linked?, European Journal of Homelessness 1 pp.201–212.

Stewart, J. (1975) Of No Fixed Abode (Manchester: Manchester University Press).

Tsemberis, S. (1999) From Streets to Homes: An Innovative Approach to Supported Housing for Homeless Adults with Psychiatric Disabilities, Journal of Community Psychology 27(2) pp.225-241.

Veit-Wilson, J.H. (1991) Paradigms of Poverty: A Rehabilitation of B.S. Rowntree Journal of Social Policy 15(1) pp.69-99.

Vorspan, R. (1977) Vagrancy and the New Poor Law in Late Victorian and Edwardian England, The English Historical Review 92(362) pp.59-81.

Williams, M. and Cheal, B. (2001) Is there Any Such Thing as Homelessness? Measurement, Explanation and Process in Homelessness Research, Innovation 14(3) pp.239-253.

Willse, C. (2010) Neo-liberal Bio-politics and the Invention of Chronic Homelessness, Economy and Society 39(2) pp.155-184. 10th Anniversary Issue 45

Homelessness Research and Policy Development: Examples from the Nordic Countries Lars Benjaminsen and Marcus Knutagård

SFI – The Danish National Centre for Social Research, Copenhagen, Denmark

School of Social Work, Lund University, Lund, Sweden

>> Abstract_ The interaction between research and policy development has played an important role in the transformation of homelessness policies and services over the last decade. Evidence from research on Housing First programmes and interventions in North America has informed the development of experimental Housing First programmes in many European countries. Moreover, studies on the profiles of homeless people and patterns of service use have supported an ongoing transition from mainly emergency and temporary responses towards strategic approaches with a focus on long-term and permanent solutions. This article will focus on the Nordic countries, where there has been close interaction between research and policy development in the formation of national strategies and programmes. In Denmark, Norway and Sweden, comprehensive national counts of homelessness have been carried out based on similar definitions and methodology and in Finland, monitoring of the extent and profiles of homelessness has been carried out for several years. In Denmark and Finland, national data have been used in the formation of national homelessness strategies that have been based on the Housing First principle, and data have been used to monitor developments in homelessness following implementation of these programmes. In Sweden, influential research on the staircase system has contributed to in-depth understanding of the unintended exclusion mechanisms that may sometimes be inherent in homelessness policies.

>> Keywords_ Homelessness research, policy development, strategic response, Housing First

ISSN 2030-2762 / ISSN 2030-3106 online 46 European Journal of Homelessness _ Volume 10, No. 3 _ 2016

Introduction

Homelessness is a complex and a persistent problem and an example of an extreme form of exclusion (Von Mahs, 2013). Considerable interaction between homelessness research and policy formation has taken place in many countries in recent years, as research has given a better understanding of the social mechanisms and dynamics behind homelessness and social exclusion, and documented the emergence of new approaches and advances in the evidence base on social interventions in the field of homelessness services.

However, it can generally be tricky to analyse the interplay between homelessness research and policy development. Policy development does not follow a linear path and research does not necessarily have a strategic focus on contributing to new policies and practices. Whilst new knowledge and evidence may be picked up in social policies and national programmes, both implementation barriers and structural barriers often mean that upscaling new social interventions and model projects and mainstreaming them into general social services is difficult.

The advanced ‘social-democratic’ welfare systems in the Nordic countries have in most cases been quite fast in picking up new trends in research and policy and after a brief overview of general trends in contemporary homelessness research, we shall use these countries as an example to study the interplay between research and policy formation in more detail. Research in the Nordic countries on both the outcomes of interventions and implementation research on policy programmes has given a deeper understanding of the mechanisms involved in these processes of knowledge diffusion and policy development.

Integrated Analytical Perspectives in the Research Literature

Recent developments in the theoretical understanding of homelessness largely follow the general synthesising trend of social theory. In the research literature, an understanding of homelessness from mainly structural or individualistic perspectives has been widely replaced by positions integrating these perspectives (Wolch and Dear, 1989; Quilgars et al., 2008; Fitzpatrick et al., 2013). From a critical realist perspective, Fitzpatrick has argued that homelessness arises from the complex interplay between structural, systemic, relational and personal factors (Fitzpatrick, 2005; 2012). According to Fitzpatrick, these factors may interact differently in both time and place, leading also to a more dynamic and contingent understanding of homelessness. Adverse structural circumstances such as unemployment or the lack of affordable housing are likely to most adversely affect people with social vulnerabilities, such as people with mental illness, substance abuse problems and weak social ties. 10th Anniversary Issue 47

The development of a more integrated and dynamic understanding of homelessness has also been expressed within the ‘pathways theory’, emphasising how homelessness should not necessarily be seen as the end point of a downward marginalisation process, as socially vulnerable individuals often experience several episodes of homelessness during their life course (Kuhn and Culhane, 1998; Shinn et al., 1998; Anderson and Tulloch, 2000; Clapham, 2005; Chamberlain and MacKenzie, 2006; Culhane et al., 2007). In their paradigmatic study on shelter data in the US, Kuhn and Culhane (1998) identified different types of shelter users: the chronically homeless with complex psychosocial problems, long spells of homelessness and long-term stays in homeless shelters; the episodically homeless with equally complex problems and many repeated spells of homelessness with a circulation in and out of shelters, prisons, hospitals and periods of rough sleeping in between; and finally the transitionally homeless – the largest group amongst US shelter users – with few, short stays in shelters and less complex problems, their homelessness likely being caused by poverty and housing affordability problems (cf. Wagner and Gilman, 2012).

A Better Understanding of Welfare Policies and Interventions

A better understanding of the impact of welfare policies has also emerged, with growing evidence that homelessness in more extensive welfare systems is more concentrated among people with psychosocial vulnerabilities, whilst homelessness in less extensive welfare systems affects wider groups of poor people due to housing affordability problems, such as the transitionally homeless in the US (Shinn, 2007; Stephens and Fitzpatrick, 2007; Toro, 2007; Benjaminsen and Andrade, 2015). As welfare systems have come under pressure in many countries, there has also been an increased focus on the unintended consequences of welfare reforms for socially vulnerable groups due to, for instance, reductions in welfare benefits or the liberalisation of housing systems (Hedin et al., 2011; Lind 2016).

In particular, changes have taken place in the understanding of homelessness interventions. In their recent volume on Housing First, Padgett et al. (2016) describe the paradigm shift from Treatment First to Housing First and how the latter has informed policy changes originating in the US and now spreading throughout Europe. By now, the Housing First approach has been incorporated into strategic responses to homelessness in large scale-national homelessness strategy programmes (Denmark and Finland) and in large-scale experimental programmes in Canada and France as well as in smaller social experimentation projects in many other EU Member States. Although debates continue over its merits in relation to particular subgroups of homeless people (Kertesz et al., 2009; see also Pleace, 2011), experimental research, mainly from the US, Canada and France, as well as evaluation research from other countries, widely documents the effectiveness of 48 European Journal of Homelessness _ Volume 10, No. 3 _ 2016

the key components of the Housing First approach: an early stabilisation of the housing situation by giving access to permanent, independent housing, and providing intensive, flexible social support following evidence-based intervention methods such as Assertive Community Treatment or Intensive Case Management, which can be differentiated according to the intensity of support needs (Tsemberis et al., 2004; Coldwell and Bendner, 2007; Nelson et al., 2007; Tsemberis, 2010).

However, in many countries, the Treatment First approach remains the predominant approach to rehousing, and research on national strategies and other large-scale Housing First programmes confirms that mainstreaming the Housing First approach into general social and homelessness services not only involves a fundamental mind-set shift in social services and the reorganisation of social services but is also hampered by major structural barriers, as shortages of affordable housing for low- income groups intensify in North American and European cities.

A Look at the Nordic Countries

The Nordic countries (Denmark, Finland, Norway and Sweden) all belong to the social-democratic welfare regime and have some of the world’s most extensive welfare systems (Esping-Andersen, 1990). Most of these countries have been relatively fast at picking up new approaches in homelessness policy and services. The most advanced examples of mainstreaming Housing First into general welfare policies are found in Denmark and Finland, where the Housing First approach has been the overall principle of large national homeless strategies. Also in Norway, recent policy development has incorporated elements of the Housing First approach. Only in Sweden, with only a few smaller ‘model projects’, has the Housing First approach not taken root to the same extent as in the other countries, likely due to the strong presence of the staircase model and stronger barriers to accessing ordinary housing for socially vulnerable people following the liberalisation of the social housing sector in Sweden.

In all four countries, we also find relatively advanced approaches to the measurement and collection of data on homelessness, and we find examples of how policy development and the monitoring of homelessness – through national homelessness counts and evaluation research on homelessness programmes – have interacted closely. This research also provides insights into the barriers and challenges to the development of integrated homeless policies even in the context of these advanced welfare states. Moreover, research shows how specific policies vary across these countries, which otherwise belong to the same welfare state cluster, leading to differences in the patterns of homelessness that can likely be linked to differences in welfare and housing policies. 10th Anniversary Issue 49

The Context of the Social Democratic Welfare System

Although the Nordic countries are generally characterised by a relatively low level of social inequality and an extensive welfare system, there are considerable differences across these countries in both social and housing policies that are important for understanding differences in the patterns and profiles of homelessness, and in homeless policies. Welfare systems in the Nordic countries, as in most other countries, are undergoing continuous reform to adapt to ongoing pressures to finance welfare services. This process started long before the international financial crisis but it has only been reinforced since then. In some cases, this has led to growing divergence in welfare systems across the Nordic countries.

Housing systems, in particular, diverge across the Nordic countries, where differences have widened due to reforms in recent decades. Norway stands out in the comparison of the countries. Norwegian housing policy is based on the homeownership ideology (Bengtsson, 2013). Whereas Denmark, Finland and Sweden have relatively large social rental sectors, Norway has a high rate of owner-occupied housing and a much lower share of social housing (Bengtsson, 2013). Moreover, the social housing sectors have been developing in different directions. Even though we can see a strengthened homeownership policy even in countries like Sweden, there is still a relatively large rental market in Denmark, Finland and Sweden. The social housing sector in Sweden, in particular, has undergone considerable reform and liberalisation, where general waiting list systems and targeted allocation systems have been widely abolished (Hedin et al., 2011; Sahlin, 2015). By contrast, the social housing sectors in Denmark and Finland still operate on a more traditional ‘social-democratic’ basis, where access is widely regulated by waiting lists and allocation systems. For example, in Denmark, municipalities have a right to allocate part of the vacancies in public housing to people in acute need of housing due to social problems. These differences set the context for homelessness policies and have broadly affected the possibilities of implementing housing- led policies, such as the Housing First approach, which has turned out to be more difficult to implement in Sweden than in Denmark or Finland.

In all Nordic countries, an accelerating process of urbanisation is occurring, with economic and population growth concentrated in capital regions and other larger urban areas. In all four countries there are signs of an increasing shortage of affordable housing in larger cities, and the provision of publicly-subsidized affordable housing – once a hallmark of the Scandinavian welfare state – is increasingly coming under pressure. 50 European Journal of Homelessness _ Volume 10, No. 3 _ 2016

Advanced Data Collection and the Monitoring of Policies

In all four countries, homelessness is monitored by national homelessness counts that provide information on the extent of homelessness and the profiles of homeless people. In Denmark, Norway and Sweden, homelessness counts are carried out using the same basic methodology, although with some variation in the definitions used. In Finland, the methodology used in homelessness enumeration is different from the other countries.

The national homelessness counts in Denmark, Norway and Sweden are conducted as an extended service based on a one-week count. The first of these counts was conducted in Sweden as early as 1993, followed by Norway in 1996, while the first Danish count was only conducted in 2007. In addition to homelessness services, a wide range of other social and health services also participate, and each unit fills out individual questionnaires for each homeless person they are in contact with or know to be homeless during the count week. Multiple use of services (double counts) are controlled for by using personal numbers or other individual information, such as initials and birth dates. As these counts are quite extensive, they are not conducted every year. In Sweden, the counts have been conducted about every five years and in Norway every third or fourth year, whereas in Denmark they have been conducted every second year since 2007.

In Finland, homelessness enumerations are conducted once every year, but the methodology is different from the other Nordic countries, as the Finnish count is an enumeration of homelessness on one single day, based on a survey administered to a wide range of homeless services and other welfare services.

Although the methodology used in the homeless counts in Denmark, Norway and Sweden is similar, there is some divergence in the definitions used. In all four countries (including Finland) it is not only rough sleepers and shelter users that are included; people staying temporarily with family or friends (couch surfers) are also defined as homeless to the extent that they are in contact with social services. The definition used in Sweden is somewhat broader than in the other countries, as it includes a wider group of people who have second-hand accommodation contracts that are often long-term, but non-permanent, whereas these groups are not included in the definitions in Denmark, Finland or Norway. This difference is not only a definitional matter but reflects the more widespread use of the staircase model in Sweden, which has led to the emergence of a secondary housing market of people who do not have ordinary leases but are in training flats, with the contract often held by social services and with behavioural conditions attached to eventually getting a permanent contract as a primary lease holder (Sahlin, 2007; Knutagård, 2009). 10th Anniversary Issue 51

In the first national count in Sweden in 1993, almost 10,000 homeless people were counted. In 1999, a new national count was done and this time the homeless numbers had dropped to 8,440. The definition of homelessness had narrowed slightly. At the end of the 1990s and the beginning of 2000 there was an ongoing debate among different actors on the causes of homelessness. Different positions were taken by different researchers and agencies. At the time, homelessness was very much seen as a result of individuals having problems with addiction and mental health. Several researchers contested this view presented by the National Board of Health and Welfare and other official agencies. This also led to a critique of the counts, especially the underlying questionnaire and the questions that were asked (Sahlin, 1996; Thörn, 2004). The critique from researchers led to a closer collaboration between the National Board of Health and Welfare and the research community. The following two national counts changed the definition once more with the ambition of being comparable with the ETHOS definition. The Swedish definition of homelessness was thereby broadened, reflecting the fact that a large share of the Swedish homeless population live as tenants in apartments sublet from the social services and dispersed in ordinary housing areas. This critique of the secondary housing market and the staircase model has been one of the most important debates in the homelessness field in Sweden in the last decade (cf. Sahlin, 1996; 2005; Löfstrand, 2005; Knutagård, 2009). The reason for including this situation in the homelessness definition is the instability and non-permanency of the situation. Tenants in the secondary housing market can easily lose their lease if they do not comply with the rules. In practice, tenants can be evicted from their apartments within a day or a week. Research also shows that there are a lot of homeless families within the secondary housing market. This leads to the risk of having children evicted from their homes. Even though the households live in apartments in ordinary housing areas, they still live in precarious situations.

Despite these variations in definition and methodology, commonalities in the approaches of Denmark, Norway and Sweden have enabled comparisons of the main trends across the countries. In 2008, a comparison of homelessness in Denmark, Norway and Sweden showed that, even when adjusting for the wider definition of homelessness in the Swedish count (and adjusting for population size), the overall level of homelessness was higher in Sweden than in Norway and Denmark. Moreover, this comparison showed that whilst the level of homelessness was of a similar size in the largest cities (Stockholm, Copenhagen, Oslo and Gothenburg), the level of homelessness in medium-sized towns was, in most cases, higher in Sweden than in similar towns in Denmark and Norway. The authors attributed this finding to a combination of the wider use of the staircase model in Sweden and the liberalisation of the Swedish social housing sector, leading to a higher level of housing exclusion even in medium-sized Swedish towns than in similar towns in Denmark and Norway (Benjaminsen and Dyb, 2008). 52 European Journal of Homelessness _ Volume 10, No. 3 _ 2016

The most recent counts in all three countries have shown that the level of homelessness is moderately increasing in Denmark, Norway and Sweden. The overall level is still higher in Sweden than in Denmark and Norway. In Denmark (population 5.7 million) the count in 2015 showed that 6,138 people were homeless – up from 4,998 in 2009, 5,290 in 2011 and 5,820 in 2013 (Benjaminsen and Lauritzen, 2015). In Norway (population 5 million), 6,259 people were homeless in 2012 – an increase from 6,091 in 2008, 5,496 in 2005 and 5,200 in 2003 (Dyb and Johannesen, 2013). The results from the 2016 count in Norway are not yet available. The recent increases in Denmark and Norway are likely attributable to a combination of factors, such as the increasing lack of affordable housing in larger cities and a particular increase in youth homelessness (Dyb and Johannesen, 2013; Rambøll and SFI, 2013; Benjaminsen and Lauritzen, 2015).

In Sweden, the latest count is from 2011 and is now five years old. It showed that about 34,000 people were homeless in a population of 9.4 million at that time. However, this strong increase compared to earlier mappings in Sweden was partly due to the widening of the definition. 13,900 people were included who were in long-term but not permanent housing, typically in the secondary housing market (Socialstyrelsen, 2012).

The methodology used in homeless enumeration is different in Finland, such that direct comparisons cannot be made between Finland and the other countries. However, the Finnish data show a very different trend, as a steady decline in homelessness has been documented over a long period; homelessness decreased from about 18,000 homeless people in 1987 – when statistics began – to about 8,000 homeless people in 2004. Since then, reductions have continued but have been relatively small. In 2014, the national statistics centre recorded 7,107 homeless people, of whom the majority were staying temporarily with family or friends (ARA survey, 2014). This can be compared to a total Finnish population of 5.5 million people in 2014. This development can widely be attributed to an intensive focus on reducing long-term homelessness in Finland, backed by a comprehensive national strategy with substantial resources devoted to establishing new housing units and converting shelters into permanent housing for long-term homeless people. Although the numbers cannot be directly compared with Denmark, Norway and Sweden, they indicate that the level of homelessness in Finland has dropped from a level more comparable to that of Sweden to a level more similar to the levels in Denmark and Norway. This may reflect Finland’s gradual move away from a mainly staircase-based approach towards the Housing First approach, which was explicitly named as the main principle in the Finnish strategy. 10th Anniversary Issue 53

Policy Developments – the Path Dependency of National Programmes

Finland was the first amongst the Nordic countries to initiate a large-scale national homelessness strategy inspired by Housing First. The Finnish programme was initiated in 2008. After Finland, Denmark was also an early starter in the introduction of Housing First, starting up a national strategy that effectively began in 2009. Sweden started up its first small-scale Housing First pilots in 2010, but it was not until 2012 that Norway started up Housing First services. By 2015, 14 municipalities had a Housing First pilot in Norway. In Sweden, the same number of pilots took five years to launch. One explanation of this development is that the start-up of Housing First projects in Norway was partly financed by The Norwegian State Housing Bank (Husbanken). In Sweden, local actors or change agents have taken on the model without any funding from the Government or other actors (Knutagård and Kristiansen, 2013).

Although previous homelessness strategies in Sweden and Norway had many goals that were similar to those of the strategies in Denmark and Finland, implementation of the strategies has turned out differently. Thus, the introduction of Housing First services is reliant on the diffusion of the model by change agents, and on their ability to translate the model into the existing system. Funding is, of course, a key ingredient. For instance, in the Finnish strategy, enough financial resources were allocated both to convert old shelters into congregate housing and to buy, renovate and build new apartments in order to reduce homelessness. The lack of financial resources from the national level has slowed down the pace of diffusion in Sweden. In the following section, we shall examine differences between the countries and the interplay between policy formation and research in more detail, discussing both commonalities and differences. However, we will narrow down the comparison to focus mainly on Denmark, Finland and Sweden.

Even though there are many similarities between the Nordic countries from a welfare regime perspective, the welfare systems differ, and the housing systems and their housing policy, in particular, differ significantly (Bengtsson, 2013). Bengtsson and his colleagues argue that the different paths the Nordic countries have chosen can be understood as examples of a path-dependent housing policy. The concept of path dependency is used to describe how early decisions will affect decisions you make later (Mahoney, 2000). In this way, the path becomes more and more difficult to break away from, since too many other parts have been added and it is therefore costly to change course. 54 European Journal of Homelessness _ Volume 10, No. 3 _ 2016

The National Homelessness Strategy in Finland – a Strategic Response

Homelessness has been a matter of concern in Finland just like the other Nordic countries. Long-term homelessness is becoming a particular concern. Over the years, the number of homeless immigrants and young people has increased. One of the more specific groups in the homelessness population in Finland is persons that have had contact with the criminal justice system; there tends to be a gap between leaving prison and being able to get help from social services with housing and other supports. The same situation applies to clients that have been in rehab or other treatment institutions. The result is often a relapse into drugs and criminal behaviour. The costs for the individual client are huge, and so is the cost for society (Pleace et al., 2016). Investing in preventative homelessness services is therefore a good strategy.

Before 2008, when Finland launched its first homelessness strategy with the explicit focus on Housing First services, the country organised most of its homelessness work in accordance with the staircase model. Homeless clients had to advance step-by-step in order to show that they were housing ready. During this period, Finland also used large hostels and other forms of temporary housing that were run down and in bad condition. This was especially true for the large-scale shelters in Helsinki.

Looking back at how homeless services have been delivered historically in Finland, the shift to a Housing First strategy is quite extraordinary. Emergency housing has almost disappeared since the 1970s. Even though it is difficult to compare the statistics of the Nordic countries, Finland is the only country that shows a decline in homelessness. In the course of the first strategy’s implementation (PAAVO I), long-term homelessness decreased by 28 percent (Pleace et al., 2015). In 2015, at the end of the second strategy (PAAVO II), for the first time fewer than 7,000 people were homeless, and long-term homelessness had dropped by 35 percent (equiva- lent to 1,345 people). The targets of the two strategies were very ambitious. The first strategy had the goal of halving long-term homelessness by 2011, while the second strategy aimed to eliminate long-term homelessness by 2015. Even though the strategy did not succeed with the goal, the reduction of long-term homelessness is still impressive in a time when pressure on the housing market is increasing. The Government’s new action plan has a focus on prevention, but also on building new dwellings. The goal is to build or allocate 2,500 dwellings by 2019 for people that are homeless or at risk of becoming homeless.1

1 www.miljoministeriet.fi/en-US/Latest_news/Shift_towards_prevention_in_reducing_hom(39553) 10th Anniversary Issue 55

The key to the reduction of homelessness in Finland is primarily the Government’s decision to convert emergency shelters into communal units. This decision made it possible to transform a temporary system quickly – a system that often became a long-term solution for homeless people, inspired by the Housing First logic. The initial conversion of shelters into communal units is one of the aspects that makes the Finnish version of Housing First a bit different to the original model. The Pathways to Housing model is based on a scattered-site principle, where apartments are scattered in the ordinary housing market in contrast to the mainly congregate housing of Finland’s first homelessness strategy. The findings showed high housing retention rates in the congregate housing too (Pleace et al, 2015). In the evaluation of the Finnish Housing First programme, scattered-site models had the highest housing retention rates, even though the congregate versions also worked for many where the treatment is separated from the housing. Even more important to the Finnish success is the focus on building, buying and renovating apartments. This has been made possible due to the financial resources invested in the strategy (Pleace et al., 2016).

The success of the strategies in Finland relies on coordination at different levels – from central government to faith-based and other organisations. The Y-foundation, which buys apartments to rent them to households in need, has been an important actor. 2 An overall conclusion is the importance of institutional entrepreneurs that have the power and position– in collaboration with other change agents – to change the mind-sets and institutional logic from a staircase logic to the principle of housing as a basic human right and a precondition for making other life changes (Hardy and Maguire, 2013; Thornton and Ocasio, 2013).

Interplay between Policy Development and Monitoring Research: the Danish National Homelessness Strategy

In Denmark, there has been considerable development in homelessness policies over the last decade. Following the first national homelessness count that took place in 2007, Denmark adopted its national homelessness strategy in 2008, with a programme period from 2009 to 2013, succeeded by a follow-up programme from 2014 to 2016 (Hansen, 2010; Benjaminsen, 2013). The strategy has been monitored at the individual level with outcome data and at an aggregate level (nationally and

2 The Y-Foundation is a major provider of social housing in Finland. The Foundation was established in 1985 by different cities in Finland and organisations such as the Association of Finnish Local and Regional Authorities, the Finnish Red Cross and the Finnish Construction Trade Union, among others. See www.ysaatio.fi/in-english/ 56 European Journal of Homelessness _ Volume 10, No. 3 _ 2016

in municipalities) with data from national homelessness counts carried out every second year since 2007. Thus, the Danish case is an example of relatively advanced interplay between policy development and data collection.

From the very beginning, the national strategy had Housing First as its overall principle, based on emerging evidence from international research on the merits of this approach. A main part of the programme was developing and testing evidence- based floating support methods (Assertive Community Treatment, Intensive Case Management and Critical Time Intervention) in municipal social services, with funding provided from the central government. Housing for the programme was widely based on public housing, where municipalities have the right to use one in four vacancies for people in acute housing need due to social problems. However, as a housing- ready approach was the norm in housing allocation policies in many municipalities, a mind-set shift was often needed to change the practice of housing allocation to the Housing First approach. Although not followed by the same rigorous experimental methods as large-scale Housing First programmes in Canada and France, an outcome monitoring system measured the situation of the approximately 1,000 people who went through the programme. The evaluation research showed that the Housing First approach was widely successful at rehousing homeless people with complex support needs, with high housing retention rates similar to those found in other Housing First projects in both Europe and the US. The strategy was succeeded by a follow-up programme aimed at mainstreaming Housing First into municipal social services, not only in the municipalities that took part in the first programme but also in new municipalities joining the programme.

Despite the ambitious national strategy programme, homelessness in Denmark has increased since 2009. In the national homelessness count in 2009, 4,998 people were recorded as homeless over the period of a week, but this number had increased to 6,138 people in 2015 (Benjaminsen and Lauritzen, 2015). The increase has been greatest in larger cities, where the shortage of affordable housing is most severe, but the most recent count also shows signs of an increase in several medium-sized cities, where a lack of affordable housing is also emerging. Count data at the municipal level reveal how transformations in homelessness patterns emerged over the period. A particularly strong increase took place in suburban municipalities in the Copenhagen area – mainly in lower- income, western suburbs that were increasingly affected by the housing shortage in the city. A strong increase was also recorded in Denmark’s second largest city, Aarhus, where homelessness more than doubled in the period. A decade ago, the level of homelessness in Aarhus was more similar to other larger provincial towns, such as Aalborg and Odense, but in 2015 the level (relative to the population size) was more similar to the level in the Copenhagen area. This transformation is likely 10th Anniversary Issue 57

a consequence of the rapid population and economic growth in the Aarhus area that has put pressure on the local housing market, with an increasing shortage of affordable housing as a consequence.

The national homelessness counts in Denmark have also documented a strong increase in youth homelessness in recent years, as the number of young homeless people (18 – 24 years of age) almost doubled from 2009 to 2015: from 633 to 1,172. The profiles show that young homeless people have largely the same high share of mental illness and substance abuse problems. About half have a mental illness and three out of four young homeless people have either a mental illness or a substance abuse problem. The count also shows that the majority of young homeless people are couch surfers staying temporarily with family and friends, whereas fewer stay in homeless shelters and very few young homeless people are sleeping rough.

The monitoring data at both the aggregated level, from the national homelessness counts, and the individual outcomes of interventions in the strategy help understand the apparently mixed results of the Danish programme. Outcome data from the national strategy show that Housing First interventions are successful for the large majority of homeless people that have received housing and support (Rambøll and SFI, 2013). However, at the same time, aggregate data at the national and municipal level from the national homelessness counts show that changes in the overall patterns of homelessness in Denmark are mainly related to the impact of more general structural transformations in society. Thus, the strategy contributed to the introduction of more effective and evidence-based interventions in social services. However, general policy and welfare reforms set the context of more specific programmes, such as the national homelessness strategy, as the increasing shortage of affordable housing and reductions in social assistance benefits for certain groups are the main barriers to scaling up and mainstreaming Housing First in general social services at the local level.

Housing and Homelessness Policies in Sweden – the Nordic Stronghold of the Staircase Model

From an optimistic perspective, it is evident that research has had an influence on how homelessness policies in Sweden have changed rhetorically over the past decade. Evidence of research results from Housing First, for example, can be seen in documents like action plans, guidelines, strategies and other policy documents at a national, regional and local level. A more pessimistic perspective can conclude, however, that homelessness research has had very little impact on homelessness policies at a national governmental level. From this perspective, Housing First initiatives can be viewed as shop window projects that exist on a small scale next to the 58 European Journal of Homelessness _ Volume 10, No. 3 _ 2016

ordinary way of organising homelessness services through the staircase model (Knutagård, 2015; Sahlin, 2015). In many ways, homelessness disappeared as a social problem in Sweden after the big Million Programme launched by the Government in the mid-1960s.3 At the end of the 1980s and the beginning of the 1990s, homelessness reappeared. At that time, homelessness was described as the ‘new homelessness’. From a Scandinavian perspective, this related, first, to the emerging phenomenon of homeless women and later to the emerging phenomena of homeless migrants and homeless families (Järvinen, 1993; 2004).

Critique of the staircase model and the secondary housing market Research on homelessness services in Sweden had shown for a long time that the staircase model had several problems. It had shown that close cooperation between social services and housing companies led to more homeless people outside the secondary housing market (Sahlin, 1996). This is a paradox, since one would assume that cooperation would lead to better services and better help for homeless people. Instead, mechanisms of exclusion were triggered. Another study showed that homeless clients were re-categorised by the services in order to be able to evict them (Löfstrand, 2005). Social services should be careful not to evict children. If the parents are divorced and the child lives with both parents off and on, the child will become homeless if one of the parents is evicted. But if the parent is re-categorised as a single household, the problem is solved on paper. Another study showed that the staircase model led to so-called institutional loops, where the homeless individual got stuck in the system of different steps to get an apartment of their own (Knutagård, 2009). One reason for this is that the individual’s housing problem was redefined as an addiction or mental health problem, so the individual had to deal with that issue first before getting a flat of their own. But, while staying at a shelter, the individual then developed new problems that demanded new services and housing options – like a boarding house, supported housing, a category house or a training flat – in order to become housing ready. The study also showed that the different housing alternatives created a rendering process, where individuals got excluded from the category house for not belonging to the right category. Even though the category house was initially available for any homeless client, the rendering process led to a redefinition of both the category house and the categories of people accepted into it.

3 The Million Programme was a large-scale public housing programme. The Social Democratic Party launched the program in 1965 and it finished in 1974. The idea was to build a million new homes in order to deal with the extreme housing shortage and poor housing standards. This very strategic response to a housing shortage was, in many ways, a very brave and radical programme. Some of the Million Programme areas have created serious problems for socio-economic segregation, but these effects were difficult to foresee. Many of the Million Programme areas, however, are very well managed (Hall and Vidén, 2005). 10th Anniversary Issue 59

The Swedish homelessness strategy Since 2002, the National Board of Health and Welfare has funded local projects to address homelessness. It was not until 2007 that the Swedish Government launched its homelessness strategy called: Homelessness: Multiple Faces, Multiple Responsibilities. The strategy had four objectives:

1. Everyone has to be guaranteed a roof over their head and be offered further coordinated action based on their individual needs.

2. The number of women and men who have been admitted to, or registered at a prison or treatment unit, or have supported accommodation or are staying in care homes and do not have any accommodation arranged before being discharged has to decrease.

3. Entry into the ordinary housing market has to be facilitated for women and men who are in housing ladders, training flats or other forms of accommodation provided by social services or other actors.

4. The number of evictions has to decrease and no children are to be evicted.

The written strategy shows that previous research has been taken into account. This is not done explicitly, but implicitly via previous reports conducted by the National Board of Health and Welfare. Three important markers are of interest here. First, the secondary housing market is used as an accepted term. The concept was coined by Professor Ingrid Sahlin of the University of Lund in the mid-1990s. Secondly, the strategy discusses the challenges and limitations with the staircase model and, finally, it announces Housing First as an alternative model that has received international attention but needs to be tried out in a Swedish context. The strategy also acknowledges that many landlords do not accept social assistance as a steady income and that this excludes people receiving assistance from getting a rental contract in the ordinary housing market. The strategy notes the need for action in order to combat this barrier.

The national strategy was evaluated, with the conclusion that the objectives of the strategy had not been met (Denvall et al., 2011). Some of the recommendations that were presented in the evaluation report were very much in line with previous research. The evaluation recommended that a new housing policy be developed with a clear focus on housing provision. Another recommendation was that a new strategy encompasses a Housing First initiative on a national level in order to try the model in a Swedish context. 60 European Journal of Homelessness _ Volume 10, No. 3 _ 2016

Housing First comes on the scene The concept of Housing First was around already at the beginning of 2000 when a public investigation presented its final report in 2001 (SOU, 2001). It was not until after Lund University held a conference on Housing First in November 2009, in order to promote the testing of the model, that a number of municipalities started up Housing First pilots. Since then, Housing First initiatives have started in around 15 municipalities. All of the pilots show housing retention rates similar to other international projects. In the municipality of Helsingborg, the housing retention rate has been up to 90 percent. The insecurity of the staircase model has been replaced by ontological security for many of the former homeless clients. The development of small Housing First services in different municipalities resembles the development in the UK but differs radically from the strategic implementation of Housing First services through national strategies in Denmark, Finland and, to some extent, Norway. An interesting result so far is that even though the pilots are quite different in their local adaptation of the Housing First model, they share the same core principles from the original Pathways to Housing model, and they seem to lead to similar outcomes.

However, looking back at how Swedish housing policy has changed, it is clear that the former idea of housing as a basic human right has been left behind in favour of housing as a commodity that the market can provide and the idea that the market can regulate itself (cf. Dorling, 2015). The research on homelessness shows that the liberalisation of Swedish housing policy has had a major impact on, and has accel- erated the housing exclusion of socially vulnerable groups. Together with the traditionally strong position of the staircase approach in Sweden, this is a major explanation of why Sweden did not see the formation of a large-scale national programme based on Housing First, like its two neighbouring Nordic countries, Denmark and Finland.

Conclusion

Research on homelessness has advanced considerably in recent years. A better understanding of the dynamics of homelessness and the interplay between structural, systemic, interpersonal and individual factors has emerged. A paradigm shift in the understanding of homelessness interventions is still taking place as the evidence on the merits of the Housing First model is growing still stronger, challenging the former paradigm of the Treatment First model. The spread of Housing First to more and more countries signifies a considerable diffusion between new research evidence and developments in practice and policy formation. 10th Anniversary Issue 61

However, research and policy developments do not always go hand in hand smoothly. Especially in a situation where an old paradigm is challenged, the interplay between research, policy development and practice in social services is complex, and the extent to which new knowledge is transformed into actual policy and practice depends on various factors. Ambiguities of research, structural barriers, institutionalised interests and barriers to changing mind-sets often explain why the relationship between new knowledge and policy formation is far from linear. To change institutionalised practises takes time, and the comparison of policy developments in the Nordic countries shows that to understand such transformations we need to take into account path dependency on previous policy developments and reforms of welfare and housing policies.

The case of the Nordic countries illustrates the interaction between research, practice and policy formation in the context of a welfare state that has been relatively open to new trends. Thus, the advanced Nordic welfare states were relatively quick to pick up new approaches to homelessness as the emerging Housing First approach started spreading from the US to Europe. Finland was the first of the Nordic countries to incorporate Housing First into general homelessness policies, and both Denmark and Finland have developed comprehensive national homelessness strategies with Housing First as the key principle. Sweden is the exception, as Housing First has not taken root there to the same extent as in the other countries, and only smaller model projects based on Housing First have emerged there. The stronghold of the staircase approach in Sweden in combination with the liberalisation of the Swedish housing sector are likely explanations for the higher barriers to incorporating Housing First into Swedish homelessness policies. Yet, despite this variation across the countries, it is increasingly challenging for municipalities in the Nordic countries to provide housing for socially vulnerable groups.

The comparison across the Nordic countries shows that the traditional social- democratic welfare state model and the high level of success of the Nordic countries in providing housing for their citizens cannot be taken for granted, as reforms of welfare and housing policies – in combination with structural factors, such as the increasing shortage of affordable housing – create new exclusion mechanisms that cannot be resolved within the domain of homelessness policies but, rather, require wider societal responses. 62 European Journal of Homelessness _ Volume 10, No. 3 _ 2016

>> References

Anderson, I. and Tulloch, D. (2000) Pathways through Homelessness: A Review of the Research Evidence (Edinburgh: Scottish Homeless Task Force).

ARA survey (2014) Homelessness in Finland 2014 (ARA — the Housing Finance and Development Centre of Finland, Report 1/2015).

Bengtsson, B. (Ed.) (2013) Varför så olika? : nordisk bostadspolitik i jämförande historiskt ljus. [Why so Different? Nordic Housing Policy in a Comparative and Historical Perspective]. (Malmö: Égalité).

Benjaminsen, L. (2013) Policy Review Up-date: Results from the Housing First based Danish Homelessness Strategy, European Journal of Homelessness 7(2) pp.109 –131.

Benjaminsen, L. and Andrade, S.B. (2015) Testing a Typology of Homelessness Across Welfare Regimes: Shelter Use in Denmark and the USA, Housing Studies 30(6) pp.858-876.

Benjaminsen, L. and Dyb, E. (2008) The Effectiveness of Homelessness Policies: Variations among the Scandinavian Countries, European Journal of Homelessness 2 pp.45 – 67.

Benjaminsen, L. and Lauritzen, H.H. (2015) Hjemløshed i Danmark 2015. National kortlægning. [Homelessness in Denmark. National Survey] (Copenhagen: SFI – The Danish National Centre for Social Research, report 15: 35).

Chamberlain, C. and MacKenzie, D. (2006) Homeless Careers: A Framework for Intervention, Australian Social Work 59(2) pp.198–212.

Clapham, D. (2005) The Meaning of Housing: A Pathways Approach (Bristol: Policy Press).

Coldwell, C. and Bendner, W. (2007) The Effectiveness of Assertive Community Treatment for Homeless Populations with Severe Mental Illness: A Meta-Analysis, American Journal of Psychiatry 164(3) pp.393–399.

Culhane, D.P., Metraux, S., Park, J.M., Schretzman, M. and Valente, J. (2007) Testing a Typology of Family Homelessness Based on Patterns of Public Shelter Utilization in Four US Jurisdictions: Implications for Policy and Program Planning, Housing Policy Debate 18(1) pp.1–27.

Denvall, V., Granlöf, S., Knutagård, M., Nordfeldt, M. and Swärd, H. (2011) Utvärdering av”Hemlöshet – många ansikten, mångas ansvar” [Evaluation of ‘Homelessness: Many Faces, Many People’s Responsibility’]. Final report (Lund: Socialhögskolan). 10th Anniversary Issue 63

Dorling, D. (2015) All that is Solid: How the Great Housing Disaster Defines our Times, and What We Can Do About It (London: Penguin Books).

Dyb, E. and Johannessen, K. (2012) Bostedsløse i Norge. En kartlegging [Homelessness in Norway: Mapping]. (Oslo: NIBR).

Esping-Andersen, G. (1990) The Three Worlds of Welfare Capitalism (Oxford: Polity Press).

Fitzpatrick, S. (2005) Explaining Homelessness: A Critical Realist Perspective, Housing, Theory and Society 22(1) pp.1–17.

Fitzpatrick, S. (2012) Homelessness, in: D. Clapham, W. Clark and K. Gibb (Eds.) The Sage Handbook of Housing Studies, pp.359–378. (London: Sage).

Fitzpatrick, S., Bramley, G. and Johnsen, S. (2013) Pathways into Multiple Exclusion: Homelessness in Seven UK Cities, Urban Studies 50(1) pp.148–168.

Hall, T. and Vidén, S. (2005) The Million Homes Programme: A Review of the Great Swedish Planning Project, Planning Perspectives 20(3) pp.301–328.

Hansen, F.K. (2010) The Homelessness Strategy in Denmark, European Journal of Homelessness 4 pp.113 –125.

Hardy, C. and Maguire, S. (2013) Institutional Entrepreneurship, in: R. Greenwood, C. Oliver, K. Sahlin and R. Suddaby (Eds.) The SAGE Handbook of Organizational Institutionalism, pp.198-217. (London: Sage).

Hedin, K., Clark, E., Lundholm, E. and Malmberg, G. (2011) Neoliberalization of Housing in Sweden: Gentrification, Filtering, and Social Polarization, Annals of the Association of American Geographers 102(2) pp.443–463.

Järvinen, M. (1993) De nye hjemløse. Kvinder, fattigdom og vold [The New Homeless: Women, Poverty and Violence]. (Holte: Forlaget socpol).

Järvinen, M. (2004) Hjemløse flygtninge og indvandrere [Homeless Refugees and Immigrants]. (Hans Reitzel: Kbh).

Kertesz, S.G., Crouch, C., Milby, J.B., Cusimano, R.E. and J. E. Schumacher (2009) Housing First for Homeless Persons with Active Addiction: Are We Overreaching?, The Milbank Quarterly 87(2) pp.495-534.

Knutagård, M. (2009) Skälens fångar. Hemlöshetsarbetets organisering, kategori- seringar och förklaringar. [Prisoners of Reasons: Organization, Categorizations and Explanations of Work with the Homeless]. Dissertation (Malmö: Égalité).

Knutagård, M. (2015) Bostad först som strategi eller strategisk strimma [Housing First as a Strategy or a Strategic Streak], Alkohol and Narkotika 4 pp.12–15. 64 European Journal of Homelessness _ Volume 10, No. 3 _ 2016

Knutagård, M. and Kristiansen, A. (2013) Not by the Book: The Emergence and Translation of Housing First in Sweden, European Journal of Homelessness 7(1) pp.93–115.

Lind, H. (2016) Åtkomliga bostäder: så gör vi det möjligt för hushåll med låga inkomster att hitta en bostad [Accessible Housing: How We Make it Possible for Low-Income Households to Find a House]. (Stockholm: SNS förlag).

Löfstrand, C. (2005) Hemlöshetens politik. Lokal policy och praktik [The Politics of Homelessness: Local Policy and Practice]. Dissertation (Malmö: Égalité).

Mahoney, J. (2000) Path Dependence in Historical Sociology, Theory and Society 29(4) pp.507-548.

Nelson, G., Aubry, T. and Lafrance, A. (2007) A Review of the Literature on the Effectiveness of Housing and Support, Assertive Community Treatment, and Intensive Case Management Interventions for Persons with Mental Illness Who Have Been Homeless, American Journal of Orthopsychiatry 77(3) pp.350–61.

Padgett, D., Henwood, B. and Tsemberis, S. (2016) Housing First: Ending Homelessness, Transforming Systems, and Changing Lives (Oxford: New York).

Pleace, N. (2011) The Ambiguities, Limits and Risks of Housing First from a European Perspective, European Journal of Homelessness 5(2) pp.113 –127.

Pleace, N., Culhane, D.P., Granfelt, R. and Knutagård, M. (2015) The Finnish Homelessness Strategy – An International Review (Helsinki: Ministry of the Environment).

Pleace, N., Knutagård, M., Culhane, D.P. and Granfelt, R. (2016) The Strategic Response to Homelessness in Finland: Exploring Innovation and Coordination within a National Plan to Reduce and Prevent Homelessness, in: N. Nichols and C. Doberstein (Eds.) Exploring Effective Systems Responses to Homelessness, pp.426-442. (Toronto: The Homeless Hub Press).

Quilgars, D., Johnsen, S. and Pleace. N. (2008) Youth Homelessness in the UK: A Decade of Progress? (York: Joseph Rowntree Foundation).

Rambøll and SFI (2013) Hjemløsestrategien. Afsluttende Rapport [The Homelessness Strategy: Final Report]. (Copenhagen: Rambøll and SFI). 10th Anniversary Issue 65

Sahlin, I. (1996) På gränsen till bostad. Avvisning, utvisning, specialkontrakt [On the Border of Housing: Rejection, Expulsion, and Special Contracts]. Dissertation (Lund: Arkiv förlag).

Sahlin, I. (2005) The Staircase of Transition: Survival through Failure, Innovation 18(2) pp.115-135.

Sahlin, I. (2007) Den sekundära bostadsmarknadens följder och förändring. En kvantitativ studie [The Impact and Change of the Secondary Housing Market: A Quantitative Study], in: C. Löfstrand Hansen and M. Nordfeldt (Eds.) Bostadslös! Lokal politik och praktik (Malmö: Gleerups).

Sahlin, I. (2015) Searching for a Homeless Strategy in Sweden, European Journal of Homelessness 9(2) pp.161–186.

Shinn, M. (2007) International Homelessness: Policy, Socio-cultural, and Individual Perspectives, Journal of Social Issues 63(3) pp.657– 677.

Shinn, M., Weitzman, B.C., Stojanovic, D. and Knickman, J.R. (1998) Predictors of Homelessness among Families in New York City: From Shelter Request to Housing Stability, American Journal of Public Health 88(1) pp.1651–1657.

Socialstyrelsen (2012) Hemlöshet och utestängning från bostadsmarknaden – Omfattning och karaktär [Homelessness and Exclusion from the Housing Market – Extent and Profile]. Stockholm: Socialstyrelsen (The National Board of Health and Welfare).

SOU (2001) Att motverka hemlöshet – en sammanhållen strategi för samhället [To Combat Homelessness – a Coherent Strategy for the Community]. (Slutbetänkande från Kommittén för hemlösa).

Stephens, M. and Fitzpatrick, S. (2007) Welfare Regimes, Housing Systems and Homelessness. How are they Linked?, European Journal of Homelessness 1 pp.201–212.

Thornton, P.H. and Ocasio, W. (2013) Institutional Logics, in: R. Greenwood, C. Oliver, K. Sahlin and R. Suddaby (Eds.) The SAGE Handbook of Organizational Institutionalism, pp.99-129. (London: Sage).

Thörn, C. (2004) Kvinnans plats(er) – bilder av hemlöshet [The Woman’s Place(s): Images of Homelessness]. Dissertation. (Stockholm: Égalité).

Toro, P.A. (2007) Toward an International Understanding of Homelessness, Journal of Social Issues 63(3) pp.461–481.

Tsemberis, S. (2010) Housing First: The Pathways Model to End Homelessness for People with Mental Illness and Addiction (Minneapolis: Hazelden Press). 66 European Journal of Homelessness _ Volume 10, No. 3 _ 2016

Tsemberis, S., Gulcur, L. and Nakae, M. (2004) Housing First, Consumer Choice and Harm Reduction for Homeless Individuals with a Dual Diagnosis, American Journal of Public Health 94(4) pp.651–656.

Von Mahs, J. (2013) Down and Out in Los Angeles and Berlin: The Sociospatial Exclusion of Homeless People (Philadelphia: Temple University Press).

Wagner, D. and Gilman, J.B. (2012) Confronting Homelessness: Poverty, Politics, and the Failure of Social Policy (Boulder, Colorado: Lynne Rienner).

Wolch, J. and Dear, M. (Eds.) (1989) The Power of Geography. How Territory Shapes Social Life (London: Unwin Hyman). 10th Anniversary Issue 67

An Emerging Research Strand: Housing Exclusion in Central and South East Europe Boróka Fehér and Nóra Teller

Budapest Social Centre and its Institutions (BMSZKI) – Budapest, Hungary

Metropolitan Research Institute – Budapest, Hungary

>> Abstract_ There was large-scale restructuring of welfare arrangements in the post-soviet states of CEE and SEE in the post-transition years, with newly emerging social challenges including various forms of housing exclusion and homelessness. In many countries, policy responses have been created ahead of evidence delivered by research, due both to the lack of a tradition of research in homelessness and to the lack of research capacity in general. In some countries, the re-criminalisation of homelessness occurred, which is the gravest example of policy formation that lacks any evidence on both reason and effect. In the past decade, however, the CEE/SEE academy has developed more and more interest in this topic, rooted either in housing research or in welfare analysis. Papers and reports have been focusing on uncovering the nature and scale of homelessness (typically through exploratory studies, both qualitative and quantitative), and also on better understanding pathways in and out of homelessness, the effectiveness of policy responses for selected groups (including issues around the criminalisation of homelessness), and the links between housing exclusion and broader welfare arrangements. This article summarises the state of research and some evidence in the CEE and SEE region. We focus on three broad areas: 1) the generation of figures on homelessness (methods and outputs of such undertakings); 2) analysis of policy interventions and strategic approaches (national and local-level policies and innovative techniques like Housing First in the region); and 3) research on the nature and dynamics of homelessness among selected subgroups like children, women, the elderly, migrants, war victims, etc. We cover Poland, Slovakia, Czech Republic, the Baltic States, Romania, Bulgaria, Serbia, Croatia and Slovenia. Some remarks on Russia’s homeless research are also included.

ISSN 2030-2762 / ISSN 2030-3106 online 68 European Journal of Homelessness _ Volume 10, No. 3 _ 2016

>> Keywords_ Emerging research, homelessness CEE/SEE, trends in homelessness, interventions

Introduction: The Foundations of ‘New’ Research on Homelessness in Central and South East Europe

The early nineties was an era of great political and economic change in the CEE and SEE region. In most countries, the transformation period went along peacefully; however, in the former Yugoslav states, it was impacted by heavy conflicts and great demographic changes. New national state formations drastically redrew the map of Central East and South East Europe.

The setting up of new institutions happened at varying paces across the region. After serious reductions in their national gross domestic product levels (especially in the former Soviet countries), the region faced only slow economic recovery with skyrocketing unemployment rates and very low activity rates (Kornai, 2006). These changes have also necessitated institutional changes in welfare arrangements and, more importantly, have created a new role for the formerly state-funded and controlled housing system (Hegedüs, 2012). Variations exist across various economies and adaptation mechanisms, but in all countries “[T]he social rental sector was largely privatized; it became a residual sector that concentrates the most vulnerable social groups. In most countries, politics (and housing policy) only recognised the need for social housing after mass privatization and economic recovery had already taken place. Programs to expand the social rental sector did not lead to a breakthrough, and its social and financial sustainability remains very weak” (Hegedüs, 2012, p.24).

Hence, housing exclusion emerged as a widening social phenomenon and at the same time it appeared as a new political issue in CEE and SEE after the transition. The most extreme form of housing exclusion – homelessness – represents a formerly largely hidden, and in some countries even forbidden ‘behaviour’ (like in the Soviet Union and Hungary). The most visible form of it, rough sleeping, was an issue that the transition governments just learned to understand around the early nineties, and they often simply left it to the equally young civil society organisations and other charitable or faith-based organisations to handle. This resulted in very selective service delivery – not only in geographical terms, but also in the sense that the groups prioritized may not have been the most needy. 10th Anniversary Issue 69

Apparently, the backlog in developing and organising homeless services also impacted the discourse on homelessness, and with this, the research on it. We can identify some distinctively different strands for ‘new’ research on homelessness in the region, like juridical studies, criminology and studies on deviances, mental health and psychological studies, and sociological studies. The juridical strand shows also some variations according to historical developments in the region; the Balkan war caused mass displacement, e.g. in Serbia and Croatia, which has prohibited many people from establishing housing security in new places or returning to their former homes. Thus, it is largely the human rights discourse that frames research on the housing exclusion of such families (Petrovic and Timotijevic, 2013). Another recent example on a juridical approach to the causes of homelessness covers all EU Member States, including those in the CEE region. This study from 2016 focuses on elements of the juridical systems of Member States and their links with broader social protection systems to understand the levels of risk to the most excluded in society in terms of losing their homes (Human European Consultancy, 2016).

Another strand approaches homelessness as an expression of deviance, and it is mostly the discourse of criminology that takes stock of the paths into and within the homeless sector. Pleace (2000) claims that this pre-twentieth century popular view on poverty and exclusion continues to impact service provision to date, and we state that this view is basically identical to the ‘official ideology’ on homelessness in the pre-transition countries in the CEE and SEE region, where it is still prevalent despite major service delivery reforms (Borbíró, 2014). More recently, the re-criminalisation of homelessness – e.g., in Hungary and Poland – delivers new impetus to this strand of research (Browarczyk, 2013; Misetics, 2013). Beyond illuminating the historical development of criminalising homeless or vagrants, Misetics (2013), for example, also works out a nuanced conflict theory-based analytical framework of the state.

Pathways into homelessness represent another core research topic in the CEE and SEE region. Obviously, psychological and mental-health related interpretations of individual paths into the growing levels of housing exclusion have a dominant role in this framework. This regularly involves looking at the interaction of homelessness with addiction or mental health issues (Doherty and Stuttaford, 2007). Studies focusing on the rough sleeper population often address this question (see, e.g., Milackova and Rochovska (2011) for Slovakia; Paksi et al. (2010) for Hungary, or Canavan et al. (2012) for an international comparison focusing on access to service provision in selected European capital cities). 70 European Journal of Homelessness _ Volume 10, No. 3 _ 2016

Similar to the developments in homelessness research in western Europe and elsewhere some two decades ago (Busch-Geertsema, 2010), we are now seeing a reorientation of research in CEE and SEE towards a more structural understanding of homelessness. Under the ‘new orthodoxy’ transformation of the political and economic systems, welfare provision and the housing system have gained a key role in interpreting issues relating to homelessness and housing exclusion, shifting the focus of research to a broader institutional context, including the privatization of housing and the commoditisation of public services (Hegedüs, 2011), comple- mented by an increasing number of qualitative and quantitative data collection exercises on homeless people.

In the following sections we focus on the increasingly interdisciplinary research findings on homelessness in the CEE and SEE region in the past 15-20 years. We focus on three broad areas: 1) generation of figures on homelessness (methods and outputs of such undertakings); 2) analysis of policy interventions and strategic approaches (national and local level policies and innovative techniques like HF; 3) research on the nature and dynamics of homelessness among selected subgroups such as children, women, the elderly, migrants, war victims and so on. We highlight selected findings from Poland, Slovakia, Czech Republic, the Baltic states, Romania, Bulgaria, Serbia, Croatia and Slovenia. Some remarks on Russia’s homeless research are also included. This review probably covers only a repre- sentative fraction of the growing corpus of research findings in CEE.

Who Are the Homeless People? Methods and Output of Explorative Studies

There is a relatively brief history of providing shelters and other services for homeless people in the CEE/SEE region. Thus, research on service provision, the nature of homelessness, the composition of the homeless population at a given point in time, or the dynamics of the population is also rather young. Therefore, in some countries we find more explorative studies. In Romania, for example, the legal definition of homelessness dates back only to 2011 (there were also studies before this date but based on another operational definition (see Nicolaescu and Mândroviceanu, 2013)), and in Slovakia, the first guidance-like document on the homeless population, who they are and how they live, is less than 10 years old (Benova, 2008). In Serbia, the ever first survey of a sub-group of homeless people is even more recent (Petrovic and Timotijevic, 2013).

In other countries, there is a more extensive history of data collection, and hence there are outputs based on a variety of research methods. This is also linked with financing research; in some countries, only very small-scale research projects get 10th Anniversary Issue 71

funded because homelessness is generally not the focus of policy makers, academia or other research organisations, or the funders of such organisations. Poland seems to have one of the longest research traditions in the region, because both qualitative and quantitative data collection started more than two decades ago. As Debski (2011) puts it: beyond local surveys mostly carried out by NGOs who run specific services, there are surveys that cover even full provinces and that reach out to homeless people who are outside of institutions. In Poland, there were attempts to count homeless people as early as the 2002 census. In 2013, there was a national point-in- time survey, and there were earlier attempts to use register-based elaborations, but the latter ones seemed difficult due to the fragmented availability of data. Despite the advanced nature of homelessness research, Busch-Geertsema et al. (2014) add that the various point in-time surveys revealed some methodological challenges in Poland, which were caused by the combination of survey data and social assistance datasets or, more recently, by combining reports from various actors like the police or individual workers, based on voluntary participation, which may have impacted both the coverage and reliability of the data produced. The Polish situation is also particularly advanced in a further sense: since 2001, biannual panel research has taken place in one of the provinces, which includes the collection of data on the spatial location of homeless people outside of institutions, as well as the service needs of the people, based on which improvements of service delivery are designed (Debski, 2011). Wygnańska (2015) points out that, based on lessons learnt from the MPHASIS project – an EU-funded project to harmonise homelessness data-collection methods – a special data collection1 was launched in 2010 in Warsaw that makes it possible to generate both stock and flow data.

Another country with a long history of data collection is Hungary, where a cross- sectional data collection has been organised on February 3 every year for the past 18 years; based on some anonymised personal information, it allows for follow-up of respondents across the datasets (Győri, 2013). Each year, recurring topical modules are added to the survey, such as on housing episodes, health status, labour market position, history of traumas, etc. One limitation of the yearly February 3 survey is that it is based on the voluntary participation of service providers, and the self-completion questionnaires may result in some inconsistencies of responses (Busch-Geertsema et al., 2014).

In response to the shortage of data on homelessness, an attempt was made in Romania to produce information based on available small-scale surveys. Dan and Dan (2005) used the 2003 housing and quality of life survey results to estimate the number of homeless people, combined with data on shelters received from 226

1 It is called BIWM, which is an acronym that comes from the Polish for ‘homelessness and housing exclusion’: bezdomność i wykluczenie mieszkaniowe. 72 European Journal of Homelessness _ Volume 10, No. 3 _ 2016

local councils (representing approximately 85% of Romania’s councils). The collected data were cleaned to avoid data bias. Then, based on cluster analysis and additional data on evictions, they extrapolated the figures for various locations in Romania. Since then, some further data estimates were produced for selected sub-groups like children (Briciu, 2014).

In Lithuania, Kanopienė and Mikulionienė (2004a and 2004b) report in detail on the results of a quantitative data collection among some 200 homeless people, on their socio-demographic characteristics and on their life trajectories. More recent quantitative data collection efforts come from Croatia, where seven large cities counted the residents of their shelters and squats in 2009 (Sostaric, 2013) and from Serbia, where the first quantitative survey since the Balkan conflict was carried out in 2011/12 in the three largest cities (Petrovic and Timotijevic, 2013).

Most recently, the 2011 census circle also delivered figures on homelessness in some countries. Based on expert reports, there are, however, some limitations to the findings from the censuses. For example, the Czech figure contained only people in overnight shelters and in homeless accommodation, and rough sleepers were not counted. The Polish figure included rough sleepers based on local government reports, yet the figure seems lower than in other counts. In Slovenia, the census category for homeless people included people living in inadequate housing situations or having a registered address at a service provider. This caused some imprecision in figures, given the fact that there is a tendency for private landlords to object to registering tenants in their properties so as to remain invisible for tax or other official purposes, such that these people are forced to register at such services. In Hungary, it was exactly the census year when homelessness was re-criminalised; thus, efforts at a complementary homeless count were stopped, especially with regard to rough sleepers or squatters (Busch-Geertsema et al., 2014).

From the available literature, we can conclude that most CEE and SEE states have a larger corpus of qualitative research on homelessness when compared to quantitative output. For example, beyond summing up research results since 2005, Briciu (2014) reports on a more recent Romanian qualitative study on clients of one of Bucharest’s service providers, which looks both at pathways into homelessness and challenges within the homeless provision system. Debski (2011) lists a whole range of reports, for example on pathways into homelessness in Poland, which is also an important topic for research in other countries, like in Slovakia where Benova (2008) describes various reasons for homelessness and people’s survival strategies, or in Croatia, where Sikic-Micanovic (2010) elaborates pathways into homelessness and how (repeated) episodes of homelessness impact the identity of homeless people. In a Czech stock-taking, Lindovska (2014) lists the estimates on and reasons for homelessness by earlier authors. Last but not least, in Russia, Stephenson (2006) 10th Anniversary Issue 73

attempts to deliver a better understanding of the process of displacement and the dead-end road that people experience, based on in-depth biographical interviews with rough sleepers. Beyond Fehér (2011), who deals with traumatic events in the life-course of Hungarian homeless adults, the Hungarian corpus is also relatively rich, because there are accompanying qualitative studies to the yearly survey on such diverse aspects as paths, strategies, housing episodes, mental dispositions and service delivery challenges (Győri and the Working Groups, 2015). Unfortunately, we find only limited research in Bulgaria to date, where Ilieva (2014) attempts to deliver some figures based on service providers and highlights some of the institutional issues related to housing exclusion based on interviews with institutional players and Hristoskova (2016) offers a thesis on the constituents of the meaning of homelessness framed by a philosophical discourse and based on a hermeneutic production of meaning. One of the first Estonian papers on the causes of homelessness, based on biographical interviews, is from 2003 (Linnas, 2003); it aims at developing guidance for social workers for more effective work with adult homeless people. In Lithuania, the research tradition is also over 15 years old; in addition to the 2003 survey, further qualitative research was carried out in more recent years on the pathways into and reasons for homelessness, the attitudes and strategies of homeless people, and the perspective of social workers on homeless people (Sadauskas, 2008).

All the above analyses deliver some insights into the emergence of and dynamic changes to homelessness in the region. The developments can partly be interpreted as identical to the general trends in western Europe. A broader understanding of the phenomenon of homelessness – beyond rough sleepers or the most visible forms of homelessness – is being developed, and analyses on the changes and restrictions to welfare arrangements, which also impact housing markets, are being published, showing how different levels of vulnerabilities among selected groups, such as young people, children, women and elderly people, contribute to homelessness.

Analyses of the Policy Environment and Strategic Approaches to Homelessness

Political and institutional transformations in the region have been feeding into research focusing on welfare arrangements and policy developments. In all countries we find descriptive papers on national and local homeless policies, the mapping of institutional arrangements and innovative service provision techniques, like housing-led or Housing First approaches. Often, such analyses deliver conclusions based on governance changes and welfare regimes discourse (Hegedüs, 2011), or they try to contextualise some of the observed changes in the housing system in general (Edgar et al., 2007; see also last section of this article). 74 European Journal of Homelessness _ Volume 10, No. 3 _ 2016

Similarly, to how qualitative research has been dealing with the question of pathways into homelessness as a core topic, there are also prominent topics in policy- oriented research reports. The largest share of institutional analyses focuses on: 1) national policies; 2) the role of different institutions in tackling or reproducing homelessness; 3) quality and cost of service delivery; and, last but not least, 4) innovations within or outside the mainstream services.

Recent examples of the analysis or review of national level strategies include Lux’s (2014) elaboration on Czech developments. He points out that the process of developing the first ever Czech national strategy on homelessness was evidence-based; while this makes it unique among strategy formulations, it is also the case that structural factors like the lack of social housing policy reform or the highly decen- tralised nature of relevant policies may hinder the implementation of the proposed measures. A slightly earlier review of the Polish strategy by Wygnańska (2009) focuses on aspects of the window of opportunity that served to accommodate innovations in the homeless service sector and sums up the challenges that were faced during the process.

Strongly linked with analyses of national policy formation are accounts of institutional and governance challenges. The emerging homelessness challenges and the backlog of reactions by states or social service providers in general seem to have resulted in very diverse solutions. For example, when comparing the Slovenian and Hungarian developments, Filipovic-Hrast et al. (2009) show that, despite very similar arrangements during pre-transition decades and the sudden emergence of visible homelessness right after or around 1990, these countries have arrived at different institutional settings and outcomes. The differences are mainly linked with varying levels of decentralisation in the two states and the scope of activity undertaken by NGOs. Whereas Slovenian service provision seems to involve a well- developed national public network, in Hungary, over two thirds of service provision is run by private players. For the Czech Republic, Hradecky (2008) points to a further issue, which is how an emerging institutional system has to deal with difficulties and tensions between ‘indigenous’ and ‘external’ service providers in meeting the needs of homeless people and dealing with selectivity in terms of clientele. Many branches of western European faith-based organisations emerged as players in the Czech Republic, and it took only short time for dedicated local volunteer organisations to become professionalised step-by-step. What seems to make the Czech case unique is that cooperation and coordination was organised in the early years by an umbrella organisation.

A more recent research strand focuses more on the day-to-day operation of the homeless provision sector. Back in 2007, a comparative analysis of UK and Poland hostel standards concluded that gaps in the quality of service delivery are closely 10th Anniversary Issue 75

related to the physical conditions of providers’ stock, and while self-regulatory mechanisms prove to be useful, for example in Poland, there should be bench- marks and monitoring in place (like in the UK) to ensure a clearer picture of what outcomes are generated in the sector, and how these are generated – especially from the perspective of users (Fitzpatrick and Wygnańska, 2007). Sostaric (2013) reviews social risk challenges of the past decades and the service responses offered to date in Croatia, and claims that despite a great step forward in 2012, when social benefits were finally – at least by law – made available for homeless people, service delivery does not fulfil its primary purpose – that is, to provide temporary accommodation until permanent housing is available. On the daily operation of the sector, she highlights the fact that there are no service standards, no coordination of activities, the competence of staff is unregulated, and there are no specialised services that enable people to leave institutions.

Accounts of the costs of homelessness from the service provision perspective are also few in number in CEE/SEE. The first analysis on costs within the sector was the FEANSTA comparative study of 2013, for which data on Poland, the Czech Republic and Hungary were collected (Pleace et al., 2013), and this fuelled further national-level research, albeit with different methodologies. For example in 2016, Radziwill (2016) reported on the cost savings of the Camillian Mission of Social Assistance compared to regular shelters in Poland, and in Hungary, Somogyi et al. (2015) compared the costs of service provision in the criminalised Hungarian service delivery context versus service delivery in housing-led projects, demon- strating that there are potential cost savings in housing-led projects.

More attention has been paid to selected projects in the field, especially since there have been larger-scale initiatives of housing-led or Housing First projects in Europe (Busch-Geertsema, 2013). Examples include Poland, where a housing reintegration project offering housing, work and health integration elements was started as early as 2004, and was followed up and monitored (Starzynsky and Wygnańska, 2007). In Hungary, in addition to short case studies published in one of the Hungarian language professional social work journals, Fehér and Balogi (2013; Balogi and Fehér, 2014) published in English on the first waves of Housing First projects, which paved the way for more sophisticated and longer-term interventions – for example, on how intensified social work assisted in putting people from the forest into private or social rental housing. On the Czech Housing First approach, Lindovska (2014) offers a critical analysis. A further recent report sums up local initiatives in Romania, Poland, Hungary and the Czech Republic on cooperation of local governments with various NGOs that house homeless people and offer floating services, or that offer preventative programmes/interventions to tackle evictions (Fehér et al., 2016). In a very recent publication looking at potential clients for Housing First projects, Wygnańska (2016) shows that serving the primary target group – that is, rough 76 European Journal of Homelessness _ Volume 10, No. 3 _ 2016

sleepers with high care needs – with scattered housing from the private rental market should be combined with various other housing options to deliver a feasible Housing First service model in Poland.

Research on the Nature and Dynamics of Homelessness Among Selected Subgroups

As research has delivered more and more evidence on the profile of homelessness in the CEE/SEE region, the dynamics of particular subgroups have become the focus of further in-depth analysis. The generic picture of the middle-aged lone man who has been rough sleeping for an average of three to six years has become more nuanced thanks to insightful analyses of youth and child homelessness (e.g., within EU-funded research projects such CSEYHP or DAPHNE, or by the WHEN network, all of which cover CEE/SEE countries) or of housing exclusion among elderly people, migrants and war victims, to list some of the groups of interest. Analyses on homeless women in the region are still emerging, and are largely framed by the domestic violence discourse (Szoboszlai, 2012).

Most of the studies apply a qualitative methodology and use in-depth life course interviews to uncover the episodes and the range of issues leading to homelessness. A European-wide comparative report of child homelessness published by FEANTSA in 2007 covers some CEE countries (European Observatory on Homelessness, 2007). A more recent report, though with a Czech national focus, is authored by Muhic-Dizdarevic and Smith (2011), who, based on expert interviews and 54 interviews with young people, explore the core issues of child and adulthood transitions that bear the risk of housing exclusion and homelessness. They also highlight the main methodological challenges of reaching out to young homeless clients/respondents, as perceived by social workers.

Migrants have been the focus of much homeless research, especially because of the dramatically low-capacity social service provision for refugees and asylum-seekers, exacerbated by the tight housing market situation. For example, in Hungary, in the post-2008 era, the visibility of homeless refugees grew. The analysis of housing pathways of migrants into and within Hungary by Kiss (2012) reveals deficiencies in service provision and the discriminatory attitudes of landlords that contribute to their homelessness. The housing exclusion of Polish migrants in western European cities is the topic of Mostowska’s research, which includes the strategies and challenges of homeless Polish people in Brussels and Oslo (Mostowska, 2011). She concludes that even in accessing low-threshold services, the lack of individual resources, especially communication skills, may interplay with legal eligibility and ultimately lead to the exclusion of migrants even from such services. 10th Anniversary Issue 77

An example of increasing attention to the needs of the elderly is presented from an international perspective by Boswell (2010), who argues that tackling isolation can be one of the core elements in the reintegration process of elderly homeless people in Poland. Debski (2010) challenges this position by showing how communities of elderly people are created and operated by one of the foundations, ultimately causing even more social isolation.

Rough sleepers are regularly the focus of research; thus, there is vast literature on the profile, housing pathways, health and mental health conditions of people living in the streets. Examples of interdisciplinary research on health, addiction and drug abuse include Paksi et al. (2010) in Hungary, who contrast the drug abuse phenomenon among Hungarian rough sleepers with the data of other European countries and of non-homeless drug addicts. They conclude that the prevalence of drug consumption among rough sleepers may be up to two or three times higher than in the general population, and that their addiction is much more entrenched than that of others who may become sober/clean sooner. There is also research on the perception of homeless people, for example by Milackova and Rochovska (2011), who analyse the historical development of homelessness in Slovakia and investigate discussions about homeless people. A further paper on the (re)presentation of homeless people in the media in Slovenia by Filipovic-Hrast (2008) claims that the representation of homeless people in the media reinforces the absence of inclusion policies or adequate housing policies.

Societal Transformations and Homelessness – Further Selected Topics in Recent Research

As already discussed above, explorative research on homelessness is more and more established throughout the region. Whereas such studies focus on the homeless population and the client groups of various services, research beyond this focus tends to centre more on structural-institutional issues that exacerbate homelessness.

In the late eighties (and even earlier in Poland and the former Yugoslavia), the loosening of state control and emerging alternative models of housing investment (self-help housing and housing cooperatives) drew increasingly on the development capacity of the general population, which contributed step-by-step to the commodification of housing in the region (Szelényi, 1989). At the same time, ‘poverty’ was taken off the political agendas of state socialist or communist parties, as it was incompatible with the social foundation of state ideology and state arrangements in general. Thus, the ‘taboo of poverty’ was revealed only after around 1989 (Romano, 2014), and it became obvious not only that policies tackling poverty were 78 European Journal of Homelessness _ Volume 10, No. 3 _ 2016

non-explicit before the political transition, but that the structures available to adapt welfare services to expanding needs were mostly ineffective. Moreover, homelessness used to be criminalised in some countries and in most large cities (Szelényi, 1996; Bence and Udvarhelyi, 2013), and its scale and scope remained, therefore, largely invisible to both the social welfare system and the general population. In addition, large groups of people effectively lost their housing with the restructuring of employment in industrial centres – for example, when workers’ hostels were closed in the Czech Republic (Hradecky, 2008) and in Hungary (Misetics, 2013).

Filipovic-Hrast et al. (2009) point out that, historically, state interventions in the homeless provision sector are launched with a backlog throughout the countries of the region, and interventions are largely based on the activities, achievements or challenges of the immature NGO sector. The slow reaction and take up of challenges by states is linked with a number of issues; in most countries, social housing and housing production were at the core of welfare provision under state socialism, and housing was perceived (and produced) as a (scarce) public good. There were housing shortages in the region, but housing amenities like water, heating and electricity were price controlled, and housing was thus affordable for the masses (Hegedüs, 2012).

After a short period of a ‘welfare gap’, when services were largely still under development, research focused on various institutional aspects of societal transformation. For example, Stephenson (2006) identified the hiatus of housing legislation, growing numbers of evictions, widespread corruption in bureaucracy, which at the same time resulted in serious housing exclusion. She also discussed patterns in the lives of homeless people, which were reminiscent of a ‘mainstream’ survival strategy. She concluded that around 2005, participation in the informal economy still played a role in maintaining some attachment with society in general, despite creating many vulnerabilities. Győri (2014) came to a similar conclusion for Hungary, exploring the drastic decrease in the labour market participation of homeless people since 2008 (from close to 50% to 7% in 2014), which also meant a dramatic decrease in available financial resources. Analyses throughout the region echo the gaps in social safety nets and the problems caused by the homeownership-based housing systems for homelessness (e.g., Edgar et al., 2007; Fehér et al., 2011; Hegedüs, 2011; Nicolaescu and Mândroviceanu, 2013; Petrovic and Timotijevic, 2013; Sostaric, 2013; Briciu, 2014). Many authors explicitly discuss the lack of financial resources as a barrier to welfare service delivery in general, whilst others are clear that more efficient use of public funds could better tackle the housing exclusion of the most vulnerable groups (Udvarhelyi, 2013), if the harmonisation of housing support programmes and benefit schemes could be designed more effectively (Fehér et al., 2011). 10th Anniversary Issue 79

Conclusion

Based on the review of paradigmatic (and available) research outputs in the central and south east European countries, we can claim that much has been accom- plished in terms of methodological developments, coverage of topics and interpre- tative frameworks, and that there is still room to do more, especially compared with the Anglo-Saxon research tradition.

We also see how much the broader European research community impacts the region’s development. Not only has there been knowledge transfer relating to research findings and methods ‘filtering’ into the region, but some policy directions have been excitedly taken up by some countries. Moreover, financial incentives in the pre-accession years to reform social service delivery, and EU-funding in general – e.g., for Housing First or housing-led approaches, have proven powerful tools in bringing the CEE/SEE community dealing with homeless research and their other European counterparts closer (see Arapoglou on policy transfers in this volume). Ferge and Juhász (2004) or Schimmelfennig and Sedelmeier (2004) made such observations already a decade ago, and more recent accounts for Poland by Wygnańska (2008) address such policy transfer mechanisms.

More comparative research on the region and across selected CEE/SEE countries could help understand some of the mechanisms leading to recurring patterns of homelessness across the countries, contextualised by converging housing systems and policies but very heterogeneous conditions of service provision. Applied research could also contribute to more effective policy formation and fuel discussions on current issues like the costs of criminalising homelessness, better profiling and diversifying service delivery for homeless people, or understanding the links between pathways out of homelessness and the constraints caused by features of the housing systems. 80 European Journal of Homelessness _ Volume 10, No. 3 _ 2016

>> References

Balogi, A. and Fehér, B. (2014) The Possibilities and Limitations of Housing-Led Projects: A Hungarian Example, European Journal of Homelessness 8(1) pp.57-75.

Bence, R. and Udvarhelyi, É.T. (2013) The Growing Criminalization of Homelessness in Hungary – A Brief Overview, European Journal of Homelessness 7(2) pp.133-143.

Benova, N. (2008) Bezdmovci, Ludia, ako my (Homeless, People Like Us) (Bratislava: Proti Prudu).

Borbíró, A. (2014) Amikor a kriminálpolitika a probléma. A hajléktalanság kriminal- izációja és az exkluzivitás veszélyei [When Criminalpolitics is the Problem. The Criminalsation of Homelessness and the danger of Exclusivity], Fundamentum 3 pp.42-45.

Boswell, C. (2010) Routes out of Poverty and Isolation for Older Homeless People: Possible Models from Poland and the UK, European Journal of Homelessness 4 pp.203-214.

Briciu, C. (2014) Homelessness in Romania – Challenges for Research and Policy, Mediterranean Journal of Social Sciences 5(22) pp.504-514.

Browarczyk, L. (2013) The Criminalisation of Homelessness in Poland, in: S. Jones (Ed) Mean Streets: A Report on Criminalisation of Homelessness in Europe, pp.91-100. (Belgium: l’Imprimerie Chauveheid).

Busch-Geertsema, V. (2010) Defining and Measuring Homelessness, in: V. Busch-Geertsema, E. O’Sullivan, D. Quilgars and N. Pleace (Eds.) Homelessness. Research in Europe. Festschrift for Joe Doherty and Bill Edgar, pp.19-39. (Brussels: FEANTSA).

Busch-Geertsema, V. (2013) Housing First Europe (Bremen: GISS).

Busch-Geertsema, V., Benjaminsen, L., Filipovič Hrast, M. and Pleace, N. (2014) Extent and Profile of Homelessness in European Member States. A Statistical Update. EOH Comparative Studies on Homelessness (Brussels: FEANTSA).

Canavan, R., Barry, M.M., Matanov, A., Barros, H., Gabor, E., Greacen, T., Holcnerova, P., Kluge, U., Nicaise, P., Moskalewicz, J., Diaz-Olalla, J.M., Straßmayr, C., Schene, A.H., Soares J.J.F., Gaddini, A., and Priebe, S. (2012) Service Provision and Barriers to Care for Homeless People with Mental Health Problems across 14 European Capital Cities, BMC Health Services Research 12(222) DOI: 10.1186/1472-6963-12-222 10th Anniversary Issue 81

Dan, A. and Dan, M-A. (2005) Persoanele fără adăpost din România – O estimare a numărului acestora [An Estimate of the Number of Homeless People in Romania], Calitatea Vieţii Xvi(1–2) pp.101–122.

Debski, M. (2010) The Existence of Communities of Homeless Persons in the Process of Vocational and Social Reintegration: Controversial Issues, European Journal of Homelessness 4 pp.261-271.

Debski, M. (2011) Homelessness in Poland: An Analysis of the Pomeranian Province, European Journal of Homelessness 5(1) pp.85 -107.

Doherty, J. and Stuttaford, M. (2007) Preventing Homelessness among Substance Users in Europe, The Journal of Primary Prevention 28(3-4) pp.245–263.

Edgar, B., Filipovic, M and Dandolova, I. (2007) Home Ownership and Marginalisation, European Journal of Homelessness 1 pp.141-160.

European Observatory on Homelessness (2007) Child Homelessness in Europe. An Overview of Emerging Trends (Brussels: FEANTSA).

Fehér, B. (2011) Understanding the Homeless Experience in Hungary through a Narrative Approach, European Journal of Homelessness 5(1) pp.49-65.

Fehér, B. and Balogi, A. (2013) From Forest to Housing: Challenges Faced by Former Rough Sleepers in the Private Rental Market in Hungary, European Journal of Homelessness 7(2) pp.63-83.

Fehér, B., Somogyi, E. and Teller, N. (2011) Támogatott lakhatási programok kiterjesztésének és más lakhatási támogatásokkal való összehangolásának lehetőségei [Supported Housing Programmes for Homeless People: Possibilities for Extension and Harmonisation with Other Housing Support Schemes in Hungary]. Esély 2012/4 pp. 6-64.

Fehér, B., Koltai, L., Németh, B., Szabó, A., Caputova, K., Scotkova, G., Tilling, J. and Wygnańska, J. (2016) Ways out of Homelessness in Central Europe. Lessons of an Erasmus Partnership (Budapest: BMSZKI).

Ferge, Z. and Juhász, G. (2004) Accession and Social Policy: The Case of Hungary, Journal of European Social Policy 14(3) pp.233-251.

Filipovic-Hrast, M. (2008) Media Representations of Homelessness and the Link to (Effective) Policies: The Case of Slovenia, European Journal of Homelessness 2 pp.115-137. 82 European Journal of Homelessness _ Volume 10, No. 3 _ 2016

Filipovic-Hrast, M., Somogyi, E. and Teller, N. (2009) The Role of NGOs in the Governance of Homelessness in Hungary and Slovenia, European Journal of Homelessness 3 pp.101-125.

Fitzpatrick, S. and Wygnańska, J. (2007) Harmonising Hostel Standards: Comparing the UK and Poland, European Journal of Homelessness 1 pp.41-66.

Győri, P. (2013) Amit tudunk – 10 év. [What we know – 10 years], in: P. Győri and J. Vida (Eds.) Otthontalanul… Tégy az emberért, pp.118-142. (Budapest: Menhely Alapítvány és BMSZKI).

Győri, P. (2014) Hajléktalan élethelyzetben lévő emberek a munkaerőpiacon [Homeless People in the Labour Market], in: Győri, P. and Vida J. (Eds.): Otthontalanul, Tégy az emberért, pp.35-44. (Budapest: BMSZKI).

Győri, P. and the Working Groups (2015) Hajléktalanügyi teendők, 2015, Utcától lakásig. Javaslatok a hajléktalan emberek ellátásának átalakítására [Obligations for the Homelessness Sector. From Street to Housing. Recommendations for the Reform of Service Provision for Homeless People]. [on-line] Available from: www. hajlekot.hu/dokumentumok/hut/hut4.0.pdf [21.12.2016].

Hegedüs, J. (2011) Social Safety Nets, the Informal Economy and Housing Poverty – Experiences in Transitional Countries, European Journal of Homelessness 5(1) pp.15-26.

Hegedüs, J. (2012) The Transformation of the Social Housing Sector in Eastern Europe. A Conceptual Framework, in: M. Lux, J. Hegedüs and N. Teller (Eds.) Social Housing in Transition Countries, pp.3-30. (New York: Routledge).

Hradecky, I. (2008) Building Capacity of Homeless Services in the Czech Republic, European Journal of Homelessness 2 pp.177-190.

Hristoskova, M.N. (2016) Philosophical Dimensions of The Problem of Homelessness. Thesis. Bulgarian Academy of Sciences

Human European Consultancy, School of Law, National University of Ireland Galway, FEANTSA (2016) Pilot project – Promoting Protection of the Right to Housing – Homelessness Prevention in the Context of Evictions. [on-line] Available from: http://ec.europa.eu/social/BlobServlet?docId=15544&langId=en [21.12.2016].

Ilieva, S. (2014) Homelessness in Bulgaria: Situation Review May 2014, Working Paper (Sofia: New Bulgarian University Scholar Electronic Repository). 10th Anniversary Issue 83

Kanopienė, V. and Mikulionienė, S. (2004a) Lietuvos benamiai: elgsena ir nuostatos [Homeless People in Lithuania: Behaviour and Attitudes], Social Work 3(1) pp.63-72.

Kanopienė, V. and Mikulionienė, S. (2004b) Lietuvos benamiai: gyvenimo sąlygos [Homeless People in Lithuania: Living Conditions], Social Work 3(1) pp.47-62.

Kiss, A. (2012) Homeless Refugees in Hungary, European Journal of Homelessness 6(2) pp.149-158.

Kornai, J. (2006) The Great Transformation of Central Eastern Europe. Success and Disappointment, Economics of Transition 14(2) pp.207–244.

Lindovska, E. (2014) Homelessness Coping Strategies from Housing Ready to Housing First Perspectives, European Journal of Homelessness 8(1) pp.97-115.

Linnas, M. (2003) Kodutus kui sotsiaalne probleem – põhjusi ja seletusi Tartu linna täisealiste kodutute näitel [Homelessness as a Social Problem – Causes and Explanations based on the Example of the Adult Homeless of Tartu], (Bakalaureusetöö (Bachelor’s Theses), Tartu Ülikool, Sotsiaalteaduskond, Tartu, 2003).

Lux, M. (2014) Milestones in the Development and Implementation of a Governmental Strategy on Homelessness in the Czech Republic, European Journal of Homelessness 8(2) pp.95-118.

Milackova, M. and Rochovska, A. (2011) Bezdomovectvo, sociálno-patologický jav vstupujúci do priestoru slovenských miest [Homelessness, Socio-pathological Phenomenon Entering the Space of Slovak Cities], Acta Geographica Universitatis Comenianae 55(2) pp.191-216.

Misetics, B. (2013) The Criminalisation of Homelessness in Hungary, in: S. Jones (Ed.) Mean Streets. A Report on Criminalisation of Homelessness in Europe, pp.101-112. (Belgium: l’Imprimerie Chauveheid).

Mostowska, M. (2011) Services for Homeless Immigrants: The Social Welfare Capital of Polish Rough Sleepers in Brussels and Oslo, European Journal of Homelessness 5(1) pp.27- 47.

Muhic-Dizdarevic, S. and Smith, J. (2011) Young Homeless People in the Czech Republic: A Comparative Perspective, European Journal of Homelessness 5(1) pp.67-83.

Nicolaescu, V. and Mândroviceanu. M. (2013) Homelessness in Romania – a Preliminary Approach, paper presented at Sustainable Ways of Preventing Homelessness conference, Copenhagen, 22 November 2013. 84 European Journal of Homelessness _ Volume 10, No. 3 _ 2016

Paksi, B., Arnold, P., Schmidt, A., Gurály, Z. and Breitner, P. (2010) Drogfogyasztás és hajléktalanság. [Drug Consumption and Homelessness], in: Győri, P. and Vida J. (Eds.): Otthontalanul… Tégy az emberért, pp.49-76. (Budapest: BMSZKI).

Petrovic, M. and Timotijevic M. (2013) Homelessness and Housing Exclusion in Serbia, European Journal of Homelessness 7(2) pp.267-286.

Pleace, N. (2000) The New Consensus, the Old Consensus and the Provision of Services for People Sleeping Rough, Housing Studies 15(4) pp.581–94.

Pleace, N., Baptista, I., Benjaminsen, L. and Busch-Geertsema, V. (2013) The Costs of Homelessness in Europe. An Assessment of the Current Evidence Base, EOH Comparative Studies on Homelessness (Brussels: FEANSTA).

Radziwill, J. (2016) Scattered Training Apartments Program and Homeless Hostel. Comparative Evaluation Based on an Example of Camillian Mission of Social Assistance, Presentation at the Conference on Housing and Support: Sustainable Way out of Homelessness, Warsaw, February 19, 2016.

Romano, S. (2014) The Political and Social Construction of Poverty: Central and Eastern European Countries in Transition (Bristol: Policy Press).

Sadauskas, J. (2008) Benamiai, kaip socialinės atskirties grupė [Social Exclusion of Homeless People], Social Work (1) pp.127-136.

Schimmelfennig, S. and Sedelmeier, U. (2004) Governance by Conditionality: EU Rule Transfer to the Candidate Countries of Central and Eastern Europe, Journal of European Public Policy 11(4) pp.661-679.

Sikic-Micanovic, L. (2010) Homelessness and Social Exclusion in Croatia, European Journal of Homelessness 4 pp.45-64.

Somogyi, E., Hegedüs, J. and Horváth, V. (2015) A kriminalizáció és hajléktalan ellátás költségeinek összevetése a támogatott lakhatás (lakhatási és szociális szolgáltatások) költségeivel [A Comparison of the Costs and Criminalization of Homeless Supported Housing (Housing and Social Services)]. (Budapest: Városkutatás Kft.).

Sostaric, M. (2013) Homelessness in the Republic of Croatia: A Review of the Social Welfare System for Homeless People, European Journal of Homelessness 7(2) pp.287-300.

Starzynski, M. and Wygnańska, J. (2007) Second Opportunity – Wola Social Reintegration Program, (Warsaw: Wola District) (in Polish, referred to in the EJH by M. Debski (2007), pp.253-255) 10th Anniversary Issue 85

Stephenson, S. (2006) Crossing the Line. Vagrancy, Homelessness and Social Displacement in Russia (Basingstoke: Ashgate).

Szelényi, I. (1989) Housing Policy in the Emergent Socialist Mixed Economy of Eastern Europe, Housing Studies 4(3) pp.167-176.

Szelényi, I. (1996) Cities under Socialism and After, in: G. Andrusz, M. Harloe and I. Szelenyi (Eds.) Cities After Socialism: Urban and Regional Change and Conflict in Post-Socialist Societies, pp.286-317. (Oxford: Blackwell).

Szoboszlai, K. (2012) Nők fedél nélkül [Roofless Women], in: Győri, P. and Vida J. (Eds.) Hajléktalan sorsok, traumák és a segítő kapcsolat. Otthontalanul… Tégy az emberért, pp.97-147. (Budapest: BMSZKI).

Udvarhelyi, É.T. (2013) (In)justice on the Streets: The Long Housing Crisis in Hungary from Above and Below (Unpublished PhD thesis, City University of New York).

Wygnańska, J. (2008) The Impact of Structural Funding on Service Provision for the Homeless (EQUAL and Poland), European Journal of Homelessness 2 pp.259-271.

Wygnańska, J. (2009) The Development of a National Homeless Strategy in Poland: Achievements and Challenges, European Journal of Homelessness 3 pp.201-214.

Wygnańska, J. (2015) Measuring Homelessness and Housing Exclusion in Poland: The BIWM Data Standard, European Journal of Homelessness 9(1) pp.123-143.

Wygnańska, J. (2016) Potential clients of the ‘Housing First’ programme in Poland. Summary of research of the ‘Housing First – Evidence based Advocacy’ Project [on-line]. Available from: www.czynajpierwmieszkanie.pl/content/ uploads/2016/04/Potential-clients-of-the-HF-programs-in-Poland_summary-of- research_wygnanska.pdf [21.12.2016]. 86 European Journal of Homelessness _ Volume 10, No. 3 _ 2016 10th Anniversary Issue 87

Researching Housing Exclusion and Homelessness in Southern Europe: Learning Through Comparing Cities and Tracking Policies Vassilis Arapoglou1

Department of Sociology, University of Crete

>> Abstract_ This article aims to highlight how cross-national research can contribute to documenting, explaining and better addressing the rising levels of housing exclusion and homelessness in southern European cities. The article introduces a historical and comparative perspective for documenting housing exclusion and homelessness, and for exploring the capacity of local policies to cope with new demands. It documents examples of inclusive housing practices and suggests that the conceptual and methodological advances of recent ‘policy mobility’ research can be useful for exploring which forms of solidarity and creativity can offer sustainable alternatives to the top-down imposition of neoliberal reforms. These observations are linked with a number of proposals for developing the research agenda on poverty and homelessness

>> Keywords_ Housing exclusion, southern Europe, policy mobility, local responses to homelessness

1 At the time of writing, the author was a senior visiting research fellow at the Hellenic Observatory and the European Institute of the London School of Economics and wishes to acknowledge its support. I wish to thank the editors of the Anniversary Issue of this Journal for their insights on the draft of this paper and especially Eoin O’ Sullivan for his detailed suggestions.

ISSN 2030-2762 / ISSN 2030-3106 online 88 European Journal of Homelessness _ Volume 10, No. 3 _ 2016

Introduction

“If we want things to stay as they are, things will have to change” (Tomasi di Lampedusa, 1960)

For more than two decades, FEANTSA has played a significant role in broadening the perspectives of its members in the countries of southern Europe and in eluci- dating the value of their own experiences for successfully linking housing with social integration policies. FEANTSA has also recently played a significant part in advancing the claims put forward by its southern European members with regard to the detrimental effects of short-sighted austerity policies. Combining advocacy and learning contributes to the politics of learning, which, as will be explained in the sections that follow, goes beyond dissemination of prefabricated solutions to local problems.

The quote above comes from the novel ‘Il Gattopardo’ (The Leopard) by Giuseppe Tomasi di Lampedusa, the last prince of Lampedusa, and it is widely cited in policy studies. It points to that fact that reforms are often of no real value: ‘change but no change’. The novel is set in Sicily in the period of the ‘Risorgimento’ – Italian unifica- tion. That Sicily and Lampedusa are today an entry point to Europe reveals the profound changes that have taken place since the decline of the Sicilian aristocracy. But history is often stubborn; it continues to shape the present and can be useful for learning.

There is another significant reason for choosing this extract. The study of the Risorgimento was influential in shaping Gramsci’s ideas about the role of hegemony – the interplay of coercion and persuasion in shaping relations between the state and civil society; how dominant classes or powerful elites use reforms to retain power, and how some radical ideas can be misinterpreted so as to become ineffective. But Gramsci was also optimistic that from the critique of “common sense”, which conceals and devalues nonconformist beliefs, “good sense” could emerge, signalling the “rough” and jagged “beginnings of the new world” (Gramsci, 1971, pp.326-343). In his philosophy of praxis, knowledge is related to creative and practical activities establishing a new worldview and a collective will for change. The conceptual framework in the following sections highlights how Gramsci’s ideas about a non-conformist and practical view of knowledge are congruent with recent ‘policy mobility’ research, especially when applied to the field of homelessness and antipoverty policies.

Clarifications about the intentions of this article are included in the introduction, because the questioning of local responses to poverty and homelessness in southern Europe could be easily misunderstood and treated in isolation from changes elsewhere in Europe. Critical inquiry is not to be confused with stereo- 10th Anniversary Issue 89

typical representations, which, in the context of the management of the crisis, have been used to blame the societies of the European south. The promoters of neoliberal welfare reforms advance talk about corruption, lack of responsibility and bureaucratic malfunction in southern European countries to justify their own failures and to block alternative pathways for change. There is another side to the story that I wish to shed some light on – namely, how distinct Mediterranean features associated with urban cultures, some of which preceded the birth of charity or state welfare in industrial cities, can sustain change. Such features include civic pride, pluralism, associationalism, municipal and local democracy movements, informal and spontaneous acts of solidarity and organised forms of community care.

The article is structured as follows. Section one provides a brief conceptual overview of the policy mobility literature. In section two there is a discussion on the contribution of policy mobility to homelessness research, highlighting the importance of comparative studies in revealing the complex links between poverty and homelessness. Section three reviews recent evidence on rapidly spreading forms of exclusion, and identifies three policy areas of special interest to southern Europe: preventing the loss of housing and making housing affordable, supported housing, and the housing of recently arrived immigrants. Section four examines the contradictions of homelessness strategies and poverty policies in the context of austerity, and discusses how some of these tensions may be addressed through a considered process of ‘translation’ and a ‘politics of learning’. The Conclusion summarises the potential of policy mobility approaches for comparing initiatives within and across different types of welfare states.

Why Learning from Difference Matters: A Brief Review of the Policy Mobility Literature

In the contemporary jargon of European institutions, much hope is vested in social innovations, experimental social policy and evidence-based social interventions. Such discourses increasingly permeate studies of homelessness and poverty.

In recognition of the shaping of public policies beyond national boundaries, a number of theories have emphasised the processes of policy transfer, multi-level governance and policy mobility (see McCann and Ward, 2011). Despite differences, all of these approaches emphasise the role of different forms of knowledge in policy change, and renew interest in transnational and trans-local networks of learning. Significantly, examples of anti-poverty, social inclusion and housing policies have been used to elaborate rather distinct approaches to policy mobility (e.g., McFarlane, 2011; Peck, 2011; Clarke et al., 2015; Peck and Theodore, 2015; Roy and Crane, 2015). 90 European Journal of Homelessness _ Volume 10, No. 3 _ 2016

In their comprehensive review of earlier approaches to the global diffusion of policies, Simmons, Dobbin and Garrett (2007) identify four distinct theories, each emphasising different pathways to change. Coercion theories explain how fiscal conservatism or trade liberalisation is enforced in economically weak states by international financial institutions, through the imposition of sanctions or conditions for granting aid. Competition theories argue that countries compete to attract investment and to sell exports by lowering labour costs, reducing constraints on investment or reducing tariff barriers. Constructivists focus on how epistemic communities and international organisations shape policy norms to combine economic development with human rights. Learning theorists suggest that countries learn from their own experiences as well as from the policy experiments of their peers.

This classification provides a good start for discerning the competing pathways of policy change, although important modifications are necessary to cope with increasing complexity and contingency in policy-making. In contemporary policy- making, the practices of coercion, competition, progressive shaping of policy norms and learning are commingled. Indeed, the blurring of lines is, to a large extent, shaped by the combined use of coercion and consent. Likewise, tensions arise as enforced policies contradict common values of social cohesion or what has been learned from local, inclusive practices.

Peck and Theodore’s (2015) ‘Fast Policy: Experimental Statecraft at the Thresholds of Neoliberalism’ makes a significant breakthrough in the critical analysis of anti- poverty policies across the world. Their politico-economic analysis of neoliberalism accords with a Gramscian view of knowledge, focusing on how actors deal with contradiction, consent or dissent from hegemonic policy norms. The book deals with how crisis-driven reforms travel and change across the cities of North and South America. It uses the examples of ‘Conditional Cash Transfer Programs’ and ‘Participatory Budgeting’ to highlight the mutations of neoliberal policies and the negative transformations of progressive practice. Peck and Theodore stress that policy mobility is not simply about the diffusion of ‘best practice’; it also consists of experiments in different forms of statecraft. Hybrid forms emerge because some key elements of policies may change from one context to another (e.g. if programmes are conditional upon willingness to work) and are thus subject to diverse and prevailing norms with regard to assisting the poor. Translation, a concept borrowed from the sociology of knowledge, is about the continuous reshaping of policy by a set of mediating actors who redesign and implement policy in new directions.

Although Peck and Theodore stress how crisis situations establish the conditions for urgent welfare reforms and the mobility of policies, their emphasis is on the expertise and leadership of cosmopolitan technocrats, rather than the coercion mechanisms with which even the designers of policies have to comply. Conformism 10th Anniversary Issue 91

is reinforced through ‘mimesis’ – a replication of branded models, whilst ‘mutability’ is a way of challenging policy orthodoxy through the changing of model components (Peck and Theodore, 2015). More controversial significance is attached to ‘modelling’ – that is, the technical framing within which solutions are sought, because technologies can either encode or disrupt neoliberal rationalities (Peck and Theodore, 2015). Nonetheless, Peck and Theodore’s emphasis on the pragmatism of policy-making in their 2015 publication underlines the loss of the critical power of some earlier writings (e.g., Peck, 2011; 2012). Much of the pragmatic rhetoric of policy makers adheres to a positivist philosophy, which undervalues both the coercive imposition of neoliberal anti-poverty reforms related to the mechanisms of austerity, and the mutations stemming from conservative attitudes towards the poor.

It could, however, be argued that a distinction between practical knowledge and those versions of pragmatism that disregard ethical and political questions is analytically useful. To Gramsci, pragmatism, in its Anglo-Saxon inception and utilitarian orientation, is concerned with changes in the immediate reality; it is a sort of ‘experimentalism’ [sic] after direct observation (Gramsci, 1971). In contrast, practical knowledge is concerned with culture and values, the setting of ethical aims, and the formation of a new ‘mentality’ disposed to the diffusion of intellectual ‘innovations’ – in effect, a ‘passion’ for and political commitment to structural change (Gramsci, 1971). In times of crisis, ethical and utilitarian concerns may diverge and ‘educational relationships’ can be seen as the way to reorganise popular values and beliefs in order to transform the world. Gramsci’s intuition that a new language of praxis is necessary in order to reflect properly the environment in which problems are formulated is instructive for the further elaboration of the concept of policy translation (Gramsci, 1971).

Freeman (2009) reminds us that the standard use of the term ‘translation’ in English refers to the physical removal from one place to another and so the term has been associated with the carrying over of meaning from one context to another. The more contexts differ, the more meanings proliferate, which is why ethnographic work on policy translation emphasises the significance of knowledge derived from local experience. Questions about the fidelity of translation are inevitably political, especially where there is a hierarchy of contexts and an uneven distribution of power across places – characteristics revealed by ethnographic interpretations of (post) colonial and neoliberal policy-making. For example, Clarke et al. (2015) highlight the fact that social policy translations are always multiple and contested, and that technocratic discourses tend to reproduce neoliberal hegemony as opposed to more creative, locally-sensitive and open-ended processes. Similarly, McFarlane (2011) argues that translation is part of a ‘politics of learning’ from both informal arrangements and the everyday life of the poor. 92 European Journal of Homelessness _ Volume 10, No. 3 _ 2016

Analytically, then, it is vital to identify tensions that may exist between the techni- calities of reforms on the one hand and cultural norms or ethical values on the other. In the European context, literature on the welfare state is revealing of the stances that welfare bureaucracies and charities take towards the poor, while policy mobility literature focuses on the democratic tradition of community change and participation. Reflecting upon the initial coining of the term ‘translation’ by Callon (1984), I would suggest that it is useful to consider how the process of translation is initiated by ‘problematization’ – that is, establishing the necessity for conducting an experiment, engaging the interest of actors, forging roles, coordinating action and recruiting allies. Problematization entails asking why anti-poverty experiments are necessary and how they are framed in terms of effectiveness, justice and democracy. Consequently, I would further argue that it is worth exploring two competing processes of translation: one which conforms to austerity-related reforms, the other facilitating a politics of learning by combining advocacy, community participation and trans-local comparisons.

Outline of a Framework for Comparing Changes in European Cities

Research associated with the European Observatory on Homelessness has documented the decisive role of different welfare systems in shaping the patterns of homelessness and housing exclusion (e.g., Stephens and Fitzpatrick, 2007; Benjaminsen et al., 2009; Busch-Geertsema et al., 2010). It has effectively documented how local governments and civil society actors interact in the design and delivery of services for homeless people, and give concrete shape to national regulations. This research tradition, however, could be expanded so as to consider the profound effects of austerity on local communities, and the capacity of local agencies and voluntary agencies to cope with rising levels of homelessness, new needs and new demands.

A distinctive Mediterranean welfare regime, often described as a Southern model of welfare, has been identified through examining the particular ways that countries like Greece, Italy, Spain and Portugal have addressed poverty and social exclusion (Mingione, 1996; Ferrera, 1996). Within Mediterranean regimes, two features are of special importance in addressing homelessness and housing exclusion. First, the Mediterranean regime was historically established as a particular version of a conservative continental one through the selective intervention of the state in the provision of education, health and housing, reflecting the claims of social groups and the political mediation of their interests. Secondly, the family has historically been the primary provider of security and welfare. It is especially in the sphere of housing that state provision has been weakest and 10th Anniversary Issue 93

family provision strongest (Allen et al., 2004). In contrast with liberal regimes, assistance to the poor was not standardised in terms of need, willingness to work or individual responsibility but was subject to fluctuations in state and civil society relationships, and local norms and values.

On the one hand, the philanthropic ideology of religious and secular charities has, in many countries of the South, been consolidated by centralised, authoritarian political regimes, wherein poverty was seen as a problem of social order, while at the same time, the legitimacy of intervention by the central state was challenged by, or relied on delicate alliances with local patrons. On the other hand, the inadequacies of social protection and restricted housing provision were compensated for by informal solidarity and spontaneous practices, as in self-housing or the development of working class settlements, often in opposition to local authoritarian regimes (Leontidou, 1990). This is the basis for the claim that widespread housing exclusion in Mediterranean regimes is related to poverty (Tosi, 1996), and the reason that visible homelessness, before the refugee crisis, remained at moderate levels, in comparison with liberal or continental regimes, albeit displaying national variations (Fitzpatrick, 1998; Busch-Geertsema et al., 2010; Busch-Geertsema et al., 2014).

It should not, then, be surprising that recent surveys on attitudes towards welfare reveal that people in the European south assign greater value to ‘need’ and ‘solidarity’ than to ‘merit’, and that lay criticism is not addressed to the welfare state per se but to its administrative inadequacies (Toro et al., 2007; Reeskens and van Oorschot, 2013; Roosma et al., 2013). Consequently, the role of civil society in the European south is not confined to formal charities, NGOs or humanitarian assistance, but also includes grassroots organisations, a variety of local solidarity initiatives and even transnational movements (Leontidou, 2016). Many such initiatives were stimulated by the anti-austerity movements in the piazzas of Madrid or Athens, and continue to emerge today in the cities and coastal areas receiving refugees.

The development of welfare provisions – universal coverage in health and education – in the 1980s was related to the restoration of democracy in many countries of the south and to attempts at state modernisation stimulated by the prospect of European integration. However, since the early 1990s, such developments have not been adequate to address needs related to persistent levels of poverty, demographic changes and migration from, mainly, the collapsing socialist states. In the same period, as the limits to European social policy were set by the fiscal conditions of the monetary union, policies to tackle social exclusion were increasingly linked to social innovation, devolution, the deployment of local partnerships, and reforms aimed at the sustainability and efficiency of social protection systems (Kazepov, 2010). Significantly, the housing question scarcely appeared in related discussions, as it was assumed that housing markets and credit expansion could provide afford- 94 European Journal of Homelessness _ Volume 10, No. 3 _ 2016

able solutions. This assumption was dramatically proved wrong in many southern European countries when the housing bubble and crash not only laid the foundations for the debt crisis but generated massive inequalities and expansive forms of exclusion (Aalbers, 2016).

Moreover, the Great Recession and the sovereign debt crisis exposed the limitations of the modernisation discourse and the inadequacy of the narrow reforms enacted within the margins of neoliberal experimentation. A critical approach has recently been advanced by Andreotti and Mingione (2016), who recognise the inability of new forms of local welfare, third sector involvement and private services to address social fragmentation. Mingione (2014) has further argued that local welfare systems respond to discrimination and profound inequalities only when social movements promoting democracy and emancipation are involved. I am further suggesting that there is a pressing need to investigate whether local welfare systems can provide integrated responses to widespread forms of housing exclusion related to hidden poverty or if they will eventually become regimes for managing the visible poor. Such contradictory tendencies can be grasped by considering the dynamics between different spatial scales – i.e., national welfare reforms induced by supranational institutions, changes in local welfare systems, and the coping strategies and needs of the poor as expressed in the geographical spaces used for their survival. Institutional changes shape the geography and demography of poor and homeless people at different scales, whether European, national or urban. As Peck (2012) has observed, the devolution of austerity – by which he means the combined deployment of responsibilities to sub-national tiers, together with public retrenchments – highlights the need to assess its impact both on homeless people and on the many agencies involved in their assistance.

The potential for real change can be discerned by distinguishing between two courses of translation and action. On the one hand, there is a politics of conforming to austerity, which includes a repertoire of responses involving the imposition of financial constraints and devolution. These may combine compliance with neoliberal ideas – like experiments with emergency solutions – with insistence on old practices, such as welfare chauvinism and philanthropic ideas about the deserving poor. We have recently documented such tendencies in Athens (see Arapoglou, 2017; Arapoglou and Gounis, 2017). On the other hand, there is a politics of learning, which capitalises on informal solidarity, integrates community responses to the needs of the poor within local development strategies and enhances the supply of affordable and supported housing. The politics of learning both scrutinizes the austerity effects of coercion/consent on the living conditions of deprived communities and opens up opportunities for democratic experiments; it aims to mirror the effects of experimentation with austerity in an accurate way and to integrate grassroots initiatives into local and trans-local networks. 10th Anniversary Issue 95

Documenting the Spread and the Deepening of Housing Exclusion

Since its formation, the European Observatory on Homelessness has had as a priority the documentation of different forms of homelessness and housing exclusion, and the pressing of European institutions and national governments to produce reliable data. The initial attempts, despite limitations, were crucial to obtaining some estimates of the extent of homelessness in the European south and to opening up research and policy agendas (e.g. Avramov, 1999). Since then, some national governments and statistical services have been more responsive (e.g., Spain and Italy) than others (e.g., Greece). From my own experience and involvement in the first steps of the Greek Network of Housing Rights, I can attest to the negligence of the administration with regard to periodic demands from activists and researchers over more than a decade. The Spanish National Statistical Institute has been carrying out surveys of homeless people and assistance providers bi-annually since 2003. Nonetheless, there is a common trend of underestimating the total number of homeless people, even in more elaborate systems such as the Spanish one (Baptista et al., 2012; Sales, 2015).

The creation of ETHOS may be considered a cornerstone in attempts to document homelessness, and its evolution is a good illustration of co-ordinated learning both within and beyond Europe (Busch-Geertsema et al., 2016). The classification has been endorsed and applied with fair success in individual cities following the intervention of organisational members of FEANTSA. In Greece, ETHOS has been adopted by the Ministry of Welfare, although its first application in 2009 was rather unfortunate (e.g., foreign nationals were not counted) and an extensive homeless survey on the basis of its categories is still pending.

Some of the limitations of the ETHOS classification should be addressed so as to better capture significant dimensions of housing exclusion in southern Europe (suggestions by García and Brändle (2014) for Spain seem applicable to other countries as well). In terms of priority, I would recommend improvement in the following areas to enhance comparability between countries and to make use of data currently available from EU-SILC (EU Statistics on Income and Living Conditions) and Eurostat’s Urban Audit: a) expand the ‘housing insecurity’ categories related to poverty and economic hardship; b) apply common EU standards to categories of inadequate housing and housing insecurity, and identify which indicators should be prioritised; and c) enhance subcategories of houseless populations, especially with regard to the housing of immigrants and asylum seekers in diverse shelter conditions, such as reception facilities, detention or deportation centres, relocation schemes, housing squats, or informal camps. 96 European Journal of Homelessness _ Volume 10, No. 3 _ 2016

The recent work by Fondation Abbé Pierre and FEANTSA (2015) on housing exclusion makes use of EU-SILC national data, and it would be worth considering expanding this to the regional level (NUTS2). There is also potential for including further measures of housing deprivation, improving sensitivity to national variations, strengthening the validity of ranking variables included in the index of housing exclusion, and mapping housing deprivation in Europe. Both the specific and the composite indicators of housing exclusion are especially relevant to the effects of the crisis on southern European countries, and so worth briefly summarising here. Poverty, housing and social exclusion, as defined by Eurostat, have increased for all countries of the European south. The four countries noted above all fall below the median value of the composite ‘European index of housing exclusion’2: of the 28 Member States, Spain ranks 15th, Portugal 21st, Italy 23rd and Greece 28th. Housing exclusion in Spain appears to be less widespread than in the other three countries, mainly because the increase in housing cost overburden (i.e. the percentage of households paying over 40% of their income in housing costs) was moderate for the total population. Yet, this picture should be modified, considering the significant recent increases in rent, mortgage arrears and housing overburden among poor households, as well as the dramatic rise of foreclosures and evictions (Ballester et al., 2015; García-Lamarca and Kaika, 2016; Kenna et al., 2016). Greece breaks records in obtaining negative values for most individual variables and, consequently, is the country where poor people are most affected by the crisis. Significantly, in all four countries the housing overburden of the poor has increased more than the European Union average, and the gap between the poor and the non-poor has increased.

National as well as regional variations are significant for learning, since they point to inequalities within the established welfare and housing regimes. Although the homeless population is concentrated in the major metropolitan and port cities, acute forms of housing exclusion are visible in some smaller cities and rural areas that host immigrants and Roma communities. Overall, from FEANTSA country profiles3 and recent reports, a common trend appears in the four countries – namely, that homelessness increased moderately during the crisis, reaching a peak around 2013/14 (Busch-Geertsema et al, 2014; Arapoglou, Gounis and Siatitsa, 2015; FIO. PSD, 2015; ISTAT 2015, Sales, 2015). Yet, demand from the poor population swells; the voluntary and the public sector can only partially meet expressed needs, mainly due to cuts in funding. Not only have the numbers of homeless people increased,

2 The composite index combines: housing costs overburden (+40% of household income), mortgage/rental arrears, overcrowding, inability to keep house warm and severe housing deprivation (see http: //www.feantsa.org/en/report/2016/09/17/an-overview-of-housing-exclusion-in- europe for details). 3 Country profiles available at http: //www.feantsa.org/en/resources/ 10th Anniversary Issue 97

but also, disturbingly, there has been a deepening in the conditions of their exclusion, especially in the deterioration of their physical and mental health (Madianos et al., 2013; FIO.PSD, 2015; Márquez and Urraza, 2015; Sypsa et al., 2015). In addition, indices of deprivation reveal an unprecedented situation regarding the levels of insecure and inadequate housing.

The more recent arrivals of displaced populations in Europe highlight new conditions of housing exclusion throughout the places they stop, especially in Greece and Italy, which are the entry points to Europe. The peak year was 2015, when the total number of those arriving by sea alone surpassed one million, according to the UNCHR – at least 857,000 in Greece and 154,000 in Italy. Arrivals slowed down considerably in 2016 after the closure of the Balkan transit route and the EU-Turkey Agreement: approximately 171,000 in Greece and 165,000 in Italy had been recorded up to November 2016. Until now, policy attention has focused on the increasing demand for emergency measures and reception arrangements, and some relief will be given through the EU-UNCHR relocation of 160,000 people from Italy and Greece. The relocation scheme, however, is only one small step; the development of an integrated response depends on the shaping of a common European immigration and asylum policy based on solidarity with those needing international protection, as well as among EU Member States.

Translation and the Politics of Learning

In the post-2008 period, the financial crisis has turned into a sovereign debt crisis, which has particularly affected some of the weakest EU states, especially Greece, Portugal, Spain, Ireland, Cyprus and Italy, although through a different combination of factors in each country. Certainly, fiscal consolidation alone cannot put countries into steady recovery and reverse the profound inequalities it has already produced. The new model of economic governance in many of the Eurozone countries includes cuts in social benefits, a decline in workforce numbers and increases in income tax. Labour market deregulation, welfare cuts and the privatisation of public assets are the primary means of granting ‘financial’ aid.

Making use of some of the ideas from policy mobility research, it could be argued that a politics of learning could contribute to revealing the contradictions of conformist thinking and to restoring a democratic and cosmopolitan ethos in policy-making. To begin with, it is important to assess the social and spatial impact of austerity, especially as it is now failing to meet the challenges of migration and mobility arising from events in the Middle East and North Africa. The section above reviewed existing evidence as to the spread and deepening of housing exclusion, but is worth repeating that research priority should be given to documenting those 98 European Journal of Homelessness _ Volume 10, No. 3 _ 2016

areas and populations that: a) are most affected by the combined effects of the Great Recession and the housing crunch – that is, those confronting the risks of both unemployment and housing insecurity; b) are in need of health care and support to retain housing – both people on the streets and those drifting towards marginal health conditions when family support proves inadequate; and c) are the focus of recently arrived immigrants and asylum seekers.

The recent deployment of the ‘Urban Agenda’ for the EU4 opens up a new arena for knowledge exchange and experimentation on how cities can address the three priority areas outlined above. This agenda provides a favourable opportunity to examine how a community development approach can enable the problematization of policies – enrolling actors and drawing on resources to address poverty, affordable housing and the inclusion of migrants and refugees. In recent years, FEANTSA has been effectively advocating innovative preventive policies and housing-led approaches, which may be sustainable if they become part of community development strategies and are linked to inclusive planning and participatory processes (Meda, 2009). In contrast to the austerity rationale, which focuses on how to manage the demand and cost of services, community development aims at sustaining economic recovery by enhancing the supply of social infrastructure and affordable housing for a variety of at-risk groups. This approach has two main advantages. First, there is more room for relaxing some of the conditions attached to income assistance for households by enhancing social housing and finding alternative means of financing it, expanding public facilities and social infrastructure, and making use of and improving the private housing stock for renting. Second, it facilitates linking housing with community services and advancing integrative solutions and prevention. Translation processes may facilitate this aim, especially if combined with learning within deprived and diverse communities and from homeless people themselves.

At the local and community level, a politics of learning implies being attentive to the survival strategies of homeless and poor people from the very beginning of the problematization of policy experiments. This implies that community knowledge is a prerequisite for change, and so research should be directed to identifying and valuing the work of grassroots groups, homeless advocates and community leaders. There is a very long tradition in urban planning of advocacy and learning, which has increasingly been inspired by cosmopolitan visions and collaborative efforts. Studying the process of translation can give new impetus to policy and participation research. I am suggesting that translation can be understood as a process of making the needs and the capacity of homeless people visible, and of facilitating the expression of their views and feelings so they are communicable.

4 http://urbanagendaforthe.eu/ 10th Anniversary Issue 99

For example, fostering an understanding of what it takes – across diverse cultures and gender lines – for a place to feel like home, or explicating how people cope with stigma and barriers to accessing services – topics that are not necessarily captured by statistics. The work of the Barcelona ‘Network of Attention to Homeless People’ (XAPSLL) in expanding the formal collection of statistics in the city is illustrative of this, as is the involvement of FIO.PSD in Italian surveys, and it would be worth translating them into other contexts, ensuring cooperation to make the results comparable, and enhancing insights through qualitative methods.

Translation can also be understood as a process of turning ‘tacit’ knowledge into ‘codified’ knowledge and modifying tools and models in response to local and individual needs. This is exceptionally important when advocating change in terms of inclusion and assistance and when proposing alternative models. For example, supported housing schemes may vary depending on the how sensitive they are to the diverse needs of substance users, families or refugees (Pleace and Bretherton, 2012). Until recently, most US-based research has focused on quantitative assessments and on how the fidelity of Housing First applications impacts on the residen- tial stability and health of clients, and not on links with communities or pathways to inclusion (Padgett et al., 2015; Quilgars and Pleace, 2016). Recent evidence suggests that Italian translations of Housing First (which, it is claimed, deviate from the fidelity model) may well have enhanced community orientation (Granelli et al., 2014; Colombo and Saruis, 2015; Oosterlynck et al., 2016).

Equally important is the enrolment of actors, coordination and mobilisation to generate wider transformation. Coordination can be viewed in terms of creating tools and connecting knowledge from distinct disciplines, especially with a view to addressing multifaceted forms of exclusion in an integrated way. This Journal hosted an enlightening discussion on the impact of housing-led initiatives on their institutional surroundings (Volume 6, 2012: ‘Responses to the Ambiguities, Limits and Risks of Housing First’). Indeed, a criterion for distinguishing between mutations of housing-led approaches should be the extent to which they make use of social housing and community-controlled assets (Hopper and Barrow, 2003). Further research is needed to evaluate the effects of housing-led initiatives on the mix of public and private provision, mental health delivery, income assistance and conceptualisations of citizenship. Recent findings indicate that applications of Housing First in southern Europe have been constrained by the scarcity of public housing, the conditional provision of very low levels of income assistance and a ‘workfarist’ orientation of recently introduced reforms (Greenwood et al., 2013; Busch-Geertsema, 2013, Oosterlynck et al., 2016). 100 European Journal of Homelessness _ Volume 10, No. 3 _ 2016

Moreover, the history of de-institutionalisation in southern Europe, starting in Italy with Franco Basaglia in the late 1960s/early 1970s, has produced a map of widespread community care in the South which challenges perceptions about its belated development (Mental Health Europe, 2012). This history also stands in contrast to the US experience, where community approaches to homelessness developed only after the failure of de-institutionalisation. Research can reignite interest in community empowerment in order to capitalise on the knowledge accumulated by mental health reformers in setting up supported employment and housing schemes, outreach and floating services, self-advocacy, and so on.

Clearly, a politics of learning as outlined by McFarlane (2011) echoes the Gramscian imperative for ethico-political knowledge aimed at regulating society and ending the internal divisions of the ruled. This claim is, of course, related to the very popular on-going discussion of inclusionary participation and self-government in urban studies, with which a number of long-standing dilemmas are associated; for example, internal divisions today cut across social, ethnic, religious and gender lines and difficult-to-reconcile tensions exist between municipal socialists, community radicals, and charities or NGOs, which often take different positions on ‘contest’ and ‘consensus’ in policy-making. Nonetheless, it is worth exploring whether means such as urban forums have been successful in solving questions of this kind and in establishing some form of collaboration between grassroots initiatives and more formalised segments of civil society. It is also worth examining if successful means of citizen participation become institutionalised and give new shape to local statecraft. Related forms of mobilisation can be found in many Mediterranean countries, but perhaps the most illustrative examples may be taken from Spain, which has been the country hardest hit by foreclosures and reposses- sions (Garcia and Haddock, 2016). The ‘PAH’ – Plataforma de Afectados por la Hipoteca (a platform for those affected by mortgages) – has been pivotal in the formation of the ‘Barcelona en Comu’ – a citizen platform launched in June 2014 that is currently governing the city of Barcelona with a strategy for defending social justice and community rights. It should not be surprising that the strength of the movement comes both from its deep roots in the Catalan history of local administration and the more recent experience of social innovation and urban citizenship (De Weerdt and Garcia, 2016; Di Feliciantonio, 2016).

Furthermore, translation is a means of advancing transnational forms of learning and advocacy and for reversing the processes that supranational institutions and elites set in motion. FEANTSA itself is an outstanding example; indeed, its own members could explain better than the research referenced in this article how they themselves have been empowered by participating in its activities. There is further potential in strengthening links with transnational urban forums – formal ones like EUKN (European Urban Knowledge Network) and HABITACT (the European exchange 10th Anniversary Issue 101

forum on local homeless strategies) or informal ones – and in accelerating the exchange and sharing of knowledge as to how best to address the reception, relocation and integration of refugees from the Middle East. The Mediterranean Sea has historically been a passage for trade and culture, as reflected in the diasporas of port and capital cities, which have become more diverse since the end of the Cold War and now include migrants from Eastern Europe, Africa and the Middle East; they are thus privileged places for strengthening Europe’s cosmopolitan image. Under conditions of austerity, aside from closing borders and setting ‘tipping-points’ for segregation and relocation quotas, there is an urgent need to codify and transfer the knowledge accumulated over the last number of decades across the many origins and destinations of immigrants. Related research and advocacy initiatives are encouraging, but research will be needed to assess how much of the potential has been realised. Another pressing matter is investigating the translating role of international NGOs and humanitarian organisations that have recently expanded their activities along the Mediterranean coast.

Conclusion

The theoretical framing, methodological innovations and themes of policy mobility research offer potential for the exploration of national and sub-national variations in the changing demography and geography of homelessness – changes that have been difficult to identify and analyse through comparisons of welfare states. Policy mobility research also offers a more complex understanding of social policy changes than the ‘Europeanisation’ paradigm, which has tended to focus on formal venues and linear processes of policy transfer. The policy mobility literature brings cities and the democratisation of social policies to the epicentre of research. Specifically, the concept of translation highlights the significance of genuine participation in advancing policy learning.

ETHOS (the European Typology of Homelessness and Housing Exclusion), and the advancement of housing-led approaches can be considered milestones in homelessness research and policy in Europe. Could their development and prospects be understood as processes of politicised learning and translation? The main part of this article has offered some insight into this idea.

The dialogue developed in the European Journal of Homelessness has contributed to the openness of the construction of both ETHOS and Housing First in Europe. It is a matter of concern that ETHOS has, for political rather than technical reasons, not been exploited in policy-making, as many researchers would have expected it to be. In a world of fast policy-making, shortcuts might involve advancing a ‘light’ version of ETHOS, and it might be worth pursuing a demonstration of research 102 European Journal of Homelessness _ Volume 10, No. 3 _ 2016

effects by involving core cities of Europe in a common exercise (as with HABITACT). Housing First in Europe seems to have been rather successful as a policy that has travelled, and it is worth advancing research on its implementation where this is linked to the development of social housing and community building. A similar approach could be taken to advance comparative research on integrated territorial and community approaches to the housing of asylum seeker and refugees. The journey to achieving such common research frames would probably be as long and complex as those that established ETHOS and Housing First in Europe.

Let me conclude with a tale borrowed from Italian colleagues working in community health promotion (Garista et al., 2015). It is a tale that Gramsci told to praise cooperation and planning to his son in a letter from prison. A mouse drinks the milk of a child in a deprived community. The mouse regrets this when the child cries, and travels to the mountain to restore the cycle of milk production. It convinces the mountain to give up its stones so that the water mill can function and the fields can be watered to grow grass for the goat to eat and make milk. The mountain agrees to give its stones and the child, when grown, plants chestnuts, oaks and pines on its slopes. Long-term planning is not congruent with fast policy-making, but is it not worth learning from the travels of this mouse how to repair suffering and restore communities of mutual exchanges and cooperation? 10th Anniversary Issue 103

>> References

Aalbers, M.B. (2016) The Financialization of Housing: A Political Economy Approach (London: Routledge).

Allen, J., Barlow, J. Leal, J. Maloutas, T. and Padovani, L. (2004) Housing and Welfare in Southern Europe (London: Wiley-Blackwell).

Andreotti, A. and Mingione, E. (2016) Local Welfare Systems in Europe and the Economic Crisis, European Urban and Regional Studies 23(3) pp.252-266.

Arapoglou V.P. (2017, forthcoming) Are Greek Cities Exceptional? Experiments with Urban Anti-poverty Policies under Austerity, in: J. Bournakis, T. Pallivos and D.K. Christopoulos (Eds.) The Greek Debt Crisis: A Political Economy Perspective (London: Palgrave Macmillan).

Arapoglou, V.P. and Gounis, K. (2017, forthcoming) Contested Landscapes of Poverty and Homelessness in South Europe: Reflections from Athens (London: Palgrave Macmillan).

Arapoglou, V., Gounis, K. and Siatitsa, D. (2015). Revisiting the Concept of Shelterisation: Insights from Athens, Greece. European Journal of Homelessness, 9(2), 137.

Avramov D. (Ed.) (1999) Coping with Homelessness: Issues to be Tackled and Best Practices in Europe (Aldershot: Ashgate).

Ballester, A., Carmona, M., Fernández, R.D., González, A., Jiménez, J., Martínez, E…. and Viano, C. (2015) The Consequences of Evictions in Spain, in: C. Walker, L. de Irmencio and M. Serder (Eds.) Social and Psychological Dimensions of Personal Debt and the Debt Industry, pp.84-100. (London: Palgrave Macmillan).

Baptista, I., Benjaminsen, L., Pleace, N. and Busch-Geertsema, V. (2012) Counting Homeless People in the 2011 Housing and Population Census, EOH Comparative Studies on Homelessness No. 2 (Brussels: FEANTSA/EOH).

Benjaminsen, L., Dyb, E. and O’Sullivan, E. (2009) The Governance of Homelessness in Liberal and Social Democratic Welfare Regimes: National Strategies and Models of Intervention, European Journal of Homelessness 3 pp.23-51.

Busch-Geertsema, V. (2013). Housing First Europe: Final Report. Bremen/ Brussels: Housing First Europe. http: //www.habitat.hu/files/ FinalReportHousingFirstEurope.pdfBusch-Geertsema, V., 104 European Journal of Homelessness _ Volume 10, No. 3 _ 2016

Busch-Geertsema, V., Edgar, W., O’Sullivan, E. and Pleace, N. (2010) Homelessness and Homeless Policies in Europe: Lessons from Research (Brussels: FEANTSA/EOH).

Busch-Geertsema, V., Benjaminsen, L., Hrast, M.F. and Pleace, N. (2014) Extent and Profile of Homelessness in European Member States: A Statistical Update (Brussels: FEANTSA/EOH).

Busch-Geertsema, V., Culhane, D. and Fitzpatrick, S. (2016) Developing a Global Framework for Conceptualising and Measuring Homelessness, Habitat International 55 pp.124-132.

Callon, M. (1984). Some Elements of a Sociology of Translation: Domestication of the Scallops and the Fishermen of St Brieuc Bay. The Sociological Review, 32(S1), 196-233.

Clarke, J., Bainton, D., Lendvai, N. and Stubbs, P. (2015) Making Policy Move: Towards a Politics of Translation and Assemblage (Bristol: Policy Press).

Colombo, F. and Saruis, T. (2015). Housing First and the City: How do innovative projects affect local policies and urban space? Paper presented at the RC21 International Conference, Urbino, Italy 27-29 August.

De Weerdt, J. and Garcia, M. (2016) Housing Crisis: the Platform of Mortgage Victims (PAH) Movement in Barcelona and Innovations in Governance, Journal of Housing and the Built Environment 31(3) pp.471-493.

Di Feliciantonio, C. (2016) Social Movements and Alternative Housing Models: Practicing the ‘Politics of Possibilities’ in Spain, Housing, Theory and Society http: //dx.doi.org/10.1080/14036096.2016.1220421

Ferrera, M. (1996). The ‘Southern Model’ of Welfare in Social Europe, Journal of European Social Policy 6(1) pp.17-37.

FIO.PSD (2015) Communicato Stampa: “Chi e povero è sempre più povero” at http: //www.fiopsd.org/wp-content/uploads/2015/12/CS_10_dicembre_2015.pdf

Fitzpatrick, S. (1998) Homelessness in the European Union, in: M. Kleinman, W. Matznetter and M. Stephens (Eds.) European Integration and Housing Policy, pp.197–204. (London: Routledge and The Royal Institution of Chartered Surveyors).

Fondation Abbé Pierre and FEANTSA (2015) An Overview of Housing Exclusion in Europe (Brussels: FEANTSA).

Freeman, R. (2009) What is ‘Translation’?, Evidence & Policy: A Journal of Research, Debate and Practice 5(4) pp.429 - 447. 10th Anniversary Issue 105

Garcia, M. and Haddock, S.V. (2016) Housing and Community Needs and Social Innovation Responses in Times of Crisis, Journal of Housing and the Built Environment 31(3) pp.1-15.

García, O. and Brändle, G. (2014) Relevance of the Use of ETHOS in the Assessment of Housing Exclusion: Proposals for Discussion from the Spanish Case, European Journal of Homelessness 8(2) pp.191-206.

García-Lamarca, M. and Kaika, M. (2016) ‘Mortgaged lives’: the Biopolitics of Debt and Housing Financialisation, Transactions of the Institute of British Geographers 41(3) pp.313-327.

Garista, P., Sardu, C., Mereu, A., Campagna, M. and Contu, P. (2015) The Mouse Gave Life to the Mountain: Gramsci and Health Promotion, Health Promotion International 30(3) pp.746-755.

Gramsci, A. (1971). Selections from the Prison Notebooks of Antonio Gramsci: Ed. and Transl. by Quintin Hoare and Geoffrey Nowell Smith. G. Nowell-Smith, & Q. Hoare (Eds.). (London: Lawrence and Wishart).

Granelli, L., Invernizzi, G. and Marchesi, L. (2014) The Rolling Stones Project: A Housing Led Experiment in Italy, European Journal of Homelessness 8(1) pp.77-96.

Greenwood, R. M., Stefancic, A., Tsemberis, S. and Busch-Geertsema, V. (2013) Implementations of Housing First in Europe: Successes and Challenges in Maintaining Model Fidelity, American Journal of Psychiatric Rehabilitation 16(4) pp.290-312.

Hopper, K. and Barrow, S.M. (2003) Two Genealogies of Supported Housing and their Implications for Outcome Assessment, Psychiatric Services 54(1) pp.50-54

ISTAT (2015). The homeless- 2014. Full text and methodological notes http:// www.istat.it/en/files/2016/06/EN_The-Homeless.pdf?title=The+homeless+- +3+Jun+2016+-+EN_The+Homeless.pdf

Kazepov, Y. (Ed.). (2010). Rescaling Social Policies: Towards Multilevel Governance in Europe (Aldershot: Ashgate).

Kenna, P., Busch-Geertsema, V., Benjaminsen, L. and Nasarre-Aznar, S. (2016) Pilot Project–Promoting Protection of the Right to Housing–Homelessness Prevention in the Context of Evictions (European Union: European Commission, Directorate-General Employment, Social Affairs and Inclusion).

Leontidou, L. (1990). The Mediterranean City in Transition: Social Change and Urban Development (Cambridge: Cambridge University Press). 106 European Journal of Homelessness _ Volume 10, No. 3 _ 2016

Leontidou, L. (2016) Urban Social Movements in Greece: Dominant Discourses and the Reproduction of ‘Weak’ Civil Societies, in: Clarke J., Huliaras A., Sotiropoulos D. (Eds.) Austerity and the Third Sector in Greece: Civil Society at the European Frontline. pp 85-108. (Oxon: Routledge).

Madianos, M.G., Chondraki, P. and Papadimitriou, G.N. (2013) Prevalence of Psychiatric Disorders among Homeless People in the Athens Area: A Cross- sectional Study, Social Psychiatry and Psychiatric Epidemiology 48(8) pp.1225-1234.

Márquez, G.M., and Urraza, X.A. (2016) The Impact of the Crisis on Homelessness in Spain: Social Intervention in a Context of Greater Need and Fewer Resources, European Journal of Social Work 19(1) pp.31-45.

McCann, E., & Ward, K. (Eds.). (2011). Mobile Urbanism: Cities and Policymaking in the Global Age (Vol. 17). (Minneapolis: U of Minnesota Press).

McFarlane, C. (2011). Learning the City: Knowledge and Translocal Assemblage (RGS-IBG Book Series Vol. 101). (Oxford: Wiley Blackwell).

Meda, J.B. (2009) How Urban Planning Instruments Can Contribute in the Fight Against Homelessness: An International Overview of Inclusionary Housing, European Journal of Homelessness 3 pp.155-177.

Mental Health Europe (2012) Mapping Exclusion: Institutional and Community Based Services in the Mental Health Field in Europe (Brussels: MHE and Open Society Foundations).

Mingione, E. (1996) Urban Poverty in the Advanced Industrial World: Concepts, Analysis and Debates, in: Mingione, E. (Ed) Urban Poverty and the Underclass: A Reader pp.3-40. (Oxford: Wiley Blackwell).

Mingione, E. (2014) The Dynamic of Capitalism, Global Commodification and Economic Crisis: What Future for Welfare in Industrialized Societies, Social Policies 1(2) pp.205-220.

Oosterlynck, S., Novy, A., Kazepov, Y., Verschraegen, G., Saruis, T., Colombo, F. and Weinzierl, C. (2016) Towards a More Effective Governance of Socially Innovative Policies–First Insights from the Case Studies (No. 16/11) (Antwerp: Herman Deleeck Centre for Social Policy, University of Antwerp).

Padgett, D., Henwood, B. and Tsemberis, S. (2015) Housing First: Ending Homelessness, Transforming Systems, and Changing Lives (New York: Oxford University Press). 10th Anniversary Issue 107

Peck, J. (2011) Global Policy Models, Globalizing Poverty Management: International Convergence or Fast-Policy Integration?, Geography Compass 5(4) pp.165-181.

Peck, J. (2012) Austerity Urbanism: American Cities under Extreme Economy, City 16(6) pp.626-655.

Peck, J. and Theodore, N. (2015) Fast Policy: Experimental Statecraft at the Thresholds of Neoliberalism (Minnesota: University of Minnesota Press).

Pleace, N., and Bretherton, J. (2012). What do we mean by Housing First? Categorising and Critically Assessing the Housing First Movement from a European Perspective. Housing: Local Welfare and Local Markets in a Globalised World. http://eprints.whiterose.ac.uk/75119

Quilgars, D. and Pleace, N. (2016) Housing First and Social Integration: A Realistic Aim?, Social Inclusion 4(4) pp.5-15.

Reeskens, T. and van Oorschot, W. (2013) Equity, Equality, or Need? A Study of Popular Preferences for Welfare Redistribution Principles across 24 European Countries, Journal of European Public Policy 20(8) pp.1174-1195.

Roosma, F., Gelissen, J. and van Oorschot, W. (2013) The Multidimensionality of Welfare State Attitudes: A European Cross-national Study, Social Indicators Research 113(1) pp.235-255.

Roy, A. and Crane, E.S. (Eds.) (2015) Territories of Poverty: Rethinking North and South (Georgia: University of Georgia Press).

Sales, A. (2015) How Many Homeless People Live in Spain? Incomplete Sources and Impossible Predictions, European Journal of Homelessness 9(2) pp.215-231.

Simmons, B. A., Dobbin, F., & Garrett, G. (2007). The Global Diffusion of Public Policies: Social Construction, Coercion, Competition or Learning?, Annual Review of Sociology 33 pp.449-472.

Stephens, M. and Fitzpatrick, S. (2007) Welfare Regimes, Housing Systems and Homelessness: How are they Linked, European Journal of Homelessness 1 pp.201-211.

Sypsa, V., Paraskevis, D., Malliori, M., Nikolopoulos, G.K., Panopoulos, A., Kantzanou, M…. and Van De Laar, M. (2015) Homelessness and Other Risk Factors for HIV Infection in the Current Outbreak among Injecting Drug Users in Athens, Greece, American Journal of Public Health 105(1) pp.196-204.

Tomasi di Lampedusa, G. (1960) The Leopard. Translation by A. Colquhoun (New York: Pantheon). 108 European Journal of Homelessness _ Volume 10, No. 3 _ 2016

Toro, P.A., Tompsett, C.J., Lombardo, S., Philippot, P., Nachtergael, H., Galand, B., Schlienz, N., Stammel, N., Yabar, Y., Blume, M., MacKay, L. and K. Harvey (2007) Homelessness in Europe and the United States: A Comparison of Prevalence and Public Opinion, Journal of Social Issues 63(3) pp.505-524.

Tosi, A. (1996) The Excluded and the Homeless: The Social Construction of the Fight against Poverty in Europe, in: E. Mingione (Ed.) Urban Poverty and the Underclass: A Reader, pp.83-104. (London: Blackwell). 10th Anniversary Issue 109

The Potential of Linked Administrative Data for Advancing Homelessness Research and Policy Dennis Culhane

University of Pennsylvania, Philadelphia, USA

>> Abstract_ Administrative data enable planners, researchers and programme evaluators to examine service use on a population basis, longitudinally. The potential for linking administrative data across multiple systems further allows for a more comprehensive view of various subpopulations, their patterns of service use and associated costs. Because people who experience homelessness are often users of multiple systems, and are often homeless in part because of ineffective programmes and insufficient aftercare, these data may be crucial to identifying the gaps that need to be filled to prevent and reduce the duration of homelessness spells. These data can also make visible what may otherwise be hidden and understudied aspects of the homelessness problem. The use of administrative data for purposes for which they were not originally intended also raises some ethical issues that must be considered in light of their potential to be misused.

>> Keywords_ Homelessness, administrative data, integrated data systems

ISSN 2030-2762 / ISSN 2030-3106 online 110 European Journal of Homelessness _ Volume 10, No. 3 _ 2016

Introduction

Recent research from Denmark (Benjaminsen and Andrade, 2105), Wales (Welsh Government, 2015), Scotland (Hamlet, 2015) and Australia (Parsell et al., 2016) demonstrates that there is a growing interest in the use of administrative data to study homelessness. Some countries in Europe, including Denmark, Scotland, Wales and France, as well as provinces in Canada (Manitoba and Ontario) and states in Australia (New South Wales and Western Australia) have gone so far as to establish archives of administrative data to support a broad variety of population and social policy research. The use of administrative data for research is not new, but their value in the social sciences is being increasingly appreciated because of their longitudinal nature. Given that they track the services and expenditures of public agencies, they have also drawn interest from governments for their policy relevance. Concerns about privacy protections and data security that may have inhibited data access in the past are being addressed through technological innovations, standardised governance structures and secure work flow processes. As these ‘integrated data systems’ (or IDS) become more available, access to administrative data could become much more routine, and could enable a new generation of research on populations with complex needs, like those who experience homelessness. In this article, I describe how the environment for administrative data- based research is changing, consider some of the opportunities that this might present for homelessness research, and reflect on some cautions regarding the ethics of this approach.

The Emerging Context for Administrative Data-based Social Science Research

Administrative data are the records collected by organisations to track their routine ‘business’ activities, including everything from banking transactions through social service contacts to medical treatment records. They can also include registration systems, like those for vital statistics (birth and death records), disease surveillance, and the like. The population and health registries in Scandinavia have been well known and appreciated by researchers for decades. Enthusiasm for ‘open data’ has put increasing pressure on governments to push vast troves of these administrative data out to the broader public, including data on crime, real estate transactions and environmental quality measures, for example. Some think that access to these ‘big data’ and the power of new computing approaches, such as machine learning, hold great promise for advancing our understanding of the world. 10th Anniversary Issue 111

The legal protections afforded to health, human service, education and earnings records (think patient, client, student and taxpayer information) have meant that these particular administrative data have been largely excluded from these types of public release. Indeed, more often than not, legislative bodies are adding protections and restrictions for access to such data out of a growing concern about the potential breach of confidential information, the vulnerability of these data to identity thieves, and their possible abuse by commercial firms. However, the potential value of these data, particularly through their linkage of cohorts of individuals across policy and service domains, and over potentially long periods of observation, has become increasingly appreciated. The fact that these data already exist, do not rely on self-report, and are population-based has drawn interest from demographers and economists. At the same time, government surveys in many countries are suffering from lagging response rates and high costs, so administrative data are viewed as offering an appealing and more economical alternative to gauge social, health and economic trends. Social inno- vators also see the potential to use administrative data to track policy experiments at low cost and in real time, improving the speed of the ‘knowledge-to-practice’ development cycle. Given these potentialities and converging interests, it is not surprising that some governments and researchers have collaborated to figure out how to improve access to these data while strengthening data security and privacy protections.

Integrated Data Systems (IDS) have emerged as a systematic approach to the archiving of sensitive and protected administrative data in some countries, and in several states and localities in the United States (Culhane et al., 2010). The typical data in an IDS can include vital records, immunisation and disease registries, school attendance and achievement, child welfare services, juvenile and adult justice data (arrest, incarceration, probation), housing assistance (including homelessness programme use), income supplements, social services for groups with disabilities, employment and earnings, retirement status and long-term care data. The model IDS creates a secure data infrastructure, enables linkage of records belonging to the same individual over time, and establishes mechanisms for data access. An IDS also usually includes a formal governance structure with representation from the various data owners and custodians. The governing board typically establishes the policies and procedures of the IDS, reviews requests for data access, and reviews study results prior to their dissemination. This ‘systematizing’ of administrative data storage and access has, in most cases, increased protection of these sensitive data, especially where previous policies were unclear or less fully articulated.

Technology innovations have also improved the security of records and protection against unintended uses (Jones et al., 2014). The encryption of records using parallel linkage keys can mean that personal identifiers do not have to be sent to 112 European Journal of Homelessness _ Volume 10, No. 3 _ 2016

the data archive from the source agencies. Computer scientists are also working on methods for creating research datasets from the different source systems ‘on the fly’ or on a request-by-request basis, meaning that a linked data archive is not even required in order to create a research dataset (removing a single hacking target). Systematic ‘perturbation’ of the underlying data, including, for example, the spatial shifting of addresses (or longitude and latitude coordinates), can provide further protection against re-identification based on location data or other potentially identifying data, like specific dates for a given treatment and diagnosis. On the researcher’s end, remote processing of the data has made it possible to provide secure access to research datasets without ever allowing researchers to view individual records (Jones et al., 2014). Instead, researchers send statistical code to the IDS and have only statistical or aggregate results returned. Multiple forms of user authentication can include certificates and biometrics (fingerprint or eye scan) to ensure that only approved users get access to research datasets and for a prescribed period of time. Results can be reviewed for disclosure risk, and cell suppression or other rules applied before release. Even better, advances in ‘differential privacy,’ which involves the introduction of intentional error in results but within bounded statistical intervals that won’t affect interpretation, provides an automated protection against the re-engineering of data to identify individuals. In effect, the results themselves are simulated.

In sum, the emerging IDS enterprise requires a whole set of legal, ethical, technological and procedural standards to be considered and developed, which might heretofore have been unimagined. Done properly, this should lead to increased data security and privacy protections, while enabling greater access to critically important data that can inform us about social policy, the economy and human development.

The Potential Value of Administrative Data for Homelessness Research

Populations with complex needs, such as people who experience homelessness and people who have multiple disadvantages, might stand to gain the most from these efforts because their well-being is dependent on the successful integration of several agencies and service systems. Indeed, the evident failure in the integration of these services directly contributes to problems like homelessness in many cases. Research using administrative data can make it possible to create useable knowledge for the reform and reorganisation of resources that can help prevent or ameliorate such conditions. Indeed, it has been argued that IDS are most useful for examining ‘key transitions’ (Fantuzzo et al., 2015) of people across developmental stages of the life course, and of people moving from one system to another, including moving out of an institutional setting – all moments of vulner- 10th Anniversary Issue 113

ability that have been linked to the onset and duration of homelessness (Metraux et al., 2010). Below, I consider some of the ways in which homelessness policy and practice could engage in and benefit from this emerging innovation in social science research.

Establishing a data collection standard for homelessness services In order to participate in a record linkage project, it is first important that the homelessness service system has a data capture system in place for registering or tracking use of homelessness programmes. In the US and Canada, and within the EU – for example, in Ireland, Hungary, the Netherlands and Denmark (Poulin et al., 2008; Busch–Geertsema et al., 2104; Government of Canada, 2016), data standards have been established that require communities or municipalities to collect basic client-level information on the users of emergency shelter services, as well as the dates of service use. These requirements generally apply only to shelter or housing programmes that receive public funding, but participation may be encouraged by all programmes in a community, regardless of their funding sources, to enable a more comprehensive picture of the problem. An obvious limitation is that these systems don’t capture people who are not users of homelessness services.

Other countries, for example Ireland and Belgium, require people to register as homeless if they benefit from homeless services or want housing priority, and perhaps to re-certify at various intervals (Pittini et al., 2015). These systems have the benefit of identifying people who might avoid homelessness services but who want access to other benefits associated with their status. A limitation is that these systems may not track service use – specifically, programme entry and exit dates. Thus, there may be only a registration date recorded, limiting longitudinal analysis of the homelessness event(s). In either case, the recording of basic client data, including names, birthdates and national identification numbers, is critical to creating the linkage keys to other datasets.

Establishing information systems of this sort is not a simple process. Many service providers are reticent about recording client-level data for a variety of reasons, including concerns that the data may be used for purposes that are not in the best interests of the clients. In order to address these concerns, a clear set of policies and a governance process have to be established. The policies should articulate the purposes of the data collection, the permissible uses of client information, secure procedures for the sharing of client data for research or evaluation purposes, and clear restrictions on the use of the client data for any purposes not otherwise outlined in the policies. Ideally, clients should be provided with notice of these policies, afforded the opportunity to opt out of data collection, and allowed to review or edit their record. 114 European Journal of Homelessness _ Volume 10, No. 3 _ 2016

It is noteworthy that record linkage projects can be undertaken on a more ad hoc basis even without such homelessness registration or services tracking systems. A recent example comes from Brisbane, Australia. Parsell and colleagues obtained the written consent of a group of formerly homeless tenants in a supported housing project to link their health (emergency, inpatient, mental health and ambulance), criminal justice (police, prison probation and parole) and homelessness service use. Tenants were assured that their identities would not be disclosed and the information derived would not be used to contact them or affect their access to benefits. Sixty-one percent of the tenants provided consent. The records were linked by the Queensland Department of Health Data Linking Authority, and identifiers were stripped from the final dataset. Researchers were then able to compare aggregate service use and costs pre- and post-housing placement, even absent a homelessness registration system or an IDS. (The authors found that aggregate service use declined from an average of $48,217 (AUS) annually prior to housing placement to $20,788 after housing placement, more than offsetting the costs of the housing programme). So, it is possible to take advantage of administrative record linkages without a homelessness services information system, although it is not as likely to be as efficient or robust for policy analysis purposes as maintaining an on-going homelessness services record system that routinely links to an IDS.

Observational studies The first benefit of having a homelessness services tracking or registration system is its enabling of basic observational studies of the nature and dynamics of the population. The service data can enable researchers to describe the incidence and prevalence of service use in the local or national population and by discrete subgroups (i.e., by ethnicity, sex, age, household status, disability groups, geography). Epidemiological studies can be undertaken of the relative risk for homelessness programme use by these subpopulations, and of the trends observed over time. Such basic statistics can inform the required scale of potential interventions, and the target subpopulations. They can also be used to assess whether interventions are having an impact on the prevalence or duration of homelessness programme use at community level, and whether other exogenous factors (recessions, broader shifts in social welfare policies) are changing the size and composition of the population over time.

Once the homelessness data are linked with other service system data – for example, through an IDS – researchers can similarly conduct observational studies of the incidence and prevalence of contiguous and/or concurrent service system programme usage. For example, rates of admission to homelessness programmes following discharge from jails or other correctional institutions can be measured, as can admission rates among people discharged from hospitals, emergency 10th Anniversary Issue 115

rooms/urgent care programmes, detoxification programmes, or youth transitioning from foster care (see, for example, Metraux et al., 2010). In effect, the homelessness programme data, through its linkage to other service system records, can be used to monitor and evaluate the effectiveness of discharge and aftercare practices of other institutions and service systems.

Correspondingly, other service systems can examine how homelessness impacts them, with their own clients as the reference populations. For example, a community might be able to observe the impact of homelessness on the local jail or on the use of emergency medical transport services, thereby motivating collaboration that might mitigate unnecessary and costly services use while improving the well-being of people who are homeless or at risk of homelessness.

Predictive analytics: identifying and refining target populations The next phase of research that is commonly supported by administrative data is looking for risk factors and interagency service-use patterns that predict entry to or exit from homelessness. A variety of methods are used to look for subpopulations within the overall population experiencing homelessness that could be targeted for preventive interventions. These can include models designed to look for the latent structure within the population (latent class analysis, factor analysis) or theory-driven approaches that seek to disaggregate the population by prede- fined variables (cluster analysis). A recent study in Denmark (Benjaminsen and Andrade, 2015), for example, undertook such an analysis, identifying the characteristics associated with patterns of homelessness programme use (see also O’Donoghue-Hynes, 2015, for a similar analysis in Ireland). Recent approaches with ‘machine learning’ have been growing in popularity; in this approach, statistical programmes look for subgroups based on iterative, best-fitting algorithmic models that are relatively agnostic to theory. Whichever approach is taken, the usual goal of this work is to identify predisposing characteristics or service-use patterns that are associated with the onset and duration of homelessness, with the hope that these can be interrupted and homelessness averted or resolved.

Consider, for example, that homelessness prevention programmes might be challenged by sceptics who could argue that many people who apply for or receive prevention services would not become homeless in the absence of those services and that, therefore, scarce resources are wasted. Indeed, a recent randomised control trial of a homelessness prevention programme in New York City (Rolston et al., 2013) found that 92 percent of the people in the control group did not become homeless despite not receiving prevention services, compared with 96 percent of the intervention group that received prevention services. Interestingly, because the intervention cut homelessness rates in half among those served (from 8 percent among controls to 4 percent among those treated), and because the cost of shelter 116 European Journal of Homelessness _ Volume 10, No. 3 _ 2016

use is so high in New York City, the intervention still resulted in modest net savings from the intervention overall. Nevertheless, the study did confirm that the vast majority of people who apply for prevention services (more than 90 percent) are unlikely to become homeless, even without the services. Thus, improved targeting could garner better results and greater efficiency, given the high rate of ‘false positive’ cases that were served. Shinn and colleagues (2013) used assessment data and administrative record linkages to develop a prevention-screening tool that has done just that, improving the targeting of New York’s prevention programme by more than 20 percent. This screener included flags based on previous homelessness programme use, behavioural health services use and child welfare involvement, in addition to self-reported factors, such as current pregnancy. Thus, in this example, predictive analytics were used to reduce the inefficiency inherent in homelessness prevention programming, improving the argument for their importance and social value.

Given the limited resources for supported housing in the US, much effort has also been invested in assessment tools and record linkage efforts to identify those among the people who experience chronic homelessness that have the highest rates of services use or risk for premature death. The ‘Vulnerability Index’ and related tools have used direct interviews with clients, allegedly to identify those most at risk of near-term mortality (Community Solutions, 2016). Other communities have linked homelessness programme records with health and justice administrative records to identify the distribution of acute care services use (hospitals and jails) in the population, and then to set thresholds for priority need populations (i.e., the top decile of users) (Flaming et al., 2009). The ethical considerations of these rationing approaches to the scarcity problem will be discussed in a later section, but it is worth noting that administrative record linkages have been used to support the triaging of housing interventions and the prioritising of target populations (and to justify denying access to some housing interventions for some groups).

Intervention testing A further, and arguably the most mature, use of linked administrative data is for the testing of interventions to assess their effectiveness. Indeed, the most common use of linked administrative data in homelessness research historically has been to assess changes in the use of services associated with the placement of people experiencing chronic or long-term homelessness in permanent supportive housing. Such pre-post study designs have been quite common in the United States, including many ad hoc record linkage projects (done only once to evaluate a single initiative or project). In a survey done in 2008 (Culhane et al., 2008), more than 50 such studies were identified in the US. Similarly, recent studies in Australia (the Brisbane study by Parsell et al., 2016, mentioned previously) and Wales (Welsh 10th Anniversary Issue 117

Government, 2015) have looked at pre- and post-intervention service use and estimated cost offsets. Generally speaking, these studies have found that a PSH (Permanent Supported Housing) placement is associated with reduced use of services, although sometimes after an initial increase in service use as part of the process of stabilising a person in their housing. Service reductions (such as reduced hospitalisations or incarcerations) have even been found to offset fully the cost of the housing intervention for some subpopulations, particularly those with high pre-housing costs, such as people with severe mental disabilities, people who are aged, or people who have extensive arrest and jail histories. However, studies generally find that cost offsets for supported housing placements (at least as commonly structured) are more modest for people with less severe disabilities and who are non-elderly. Nevertheless, even finding some reductions in acute care service use has helped to make the moral argument for further investments in housing, as people are better served in housing than being left homeless, and they use more appropriate (and less expensive) support services, regardless of overall cost offsets or realisable savings.

Because most of these pre-post study designs are quasi-experimental, have small sample sizes and very often do not include comparison groups, concerns could be raised that selection bias has skewed results in favour of finding positive cost offsets, as tenants with disproportionately higher need are more likely to be enrolled. The limitations associated with small sample sizes and a lack of comparison groups can be overcome with larger study populations and the use of administrative records to create matched control groups, including through propensity score matching (see Culhane et al., 2002). In such studies, administrative records are used to identify groups with common homelessness service use histories, comparable demographic characteristics, similar behavioural health diagnoses and shared aggregate justice system use (jail stays). These matching efforts can strengthen the argument for the robustness of the outcome, relative to what happens to the comparison group, controlling for regression to the mean or other confounders. And in many cases, this improved and more rigorous quasi-experimental approach may be as good a standard as can be achieved scientifically while also meeting community norms for ethical research, which may preclude randomi- sation of people experiencing homelessness to interventions.

However, it has also been the case that some large-scale randomised controlled trials (RCTs) have been used to examine the effectiveness of homelessness interventions, such as PSH, targeted at people experiencing long-term or chronic homelessness. Most notably, Canada launched a large, multi-city RCT to look at the impact of PSH on homelessness, including use of services pre- and post- housing placement (Stergiopoulos et al., 2015). The study was able to use tenant interviews to track service use (self-report) – an effort that is now being replicated 118 European Journal of Homelessness _ Volume 10, No. 3 _ 2016

with linked administrative data (Hwang, 2016). The study found that service reductions and cost-offsets were significant among the high-need subpopulation, but not significant in the moderate-need subpopulation. This is broadly consistent with the results from the quasi-experimental literature in the US.

Advocates of evidence-based policy-making have also been promoting the use of linked administrative data as a cost-effective and timely way to embed randomised policy experiments in the everyday practice of public administration (Haskins and Baron, 2011; Nussle and Orszag, 2015). These advocates argue that administrative data makes it possible to conduct high-speed/low-cost RCTs as a regular business practice to improve social programmes. They point to private sector businesses, like Google and Bank of America, who conduct thousands of experiments every year by varying their products and tracking the impact in real time with administrative data. Of course, changing social policy is more complex (and sensitive) than changing a website or search algorithm. But the translational message is that the prevailing dominance of carrying out social policy experiments like prescription drug trials – with relatively small samples tracked over three to five years at great expense – is being challenged by a new model of shorter, faster knowledge development cycles made possible by embedded experiments using administrative data systems to measure outcomes.

Some Cautions on the Ethical and Scientific Use of Administrative Data

The era of ‘big data’ is promising increased knowledge at increased speed, but it is also a Brave New World for the social sciences and social policy, fraught with ethical issues as well as new (or newly significant) scientific considerations. Most of these challenges can be confronted and properly addressed, but they also must be forthrightly and transparently established. The scientific community and the social policy community are often viewed with suspicion by the general public because they are perceived as doing secretive work without proper public scrutiny, and that, consequently, people can be victimised by their elitist machina- tions. Unfortunately, far too many historical examples support these suspicions. If social science and social policy are to benefit from linked administrative data, with all the perceived hazards associated with privacy compromises and with ‘Big Brother’ watching, public confidence and trust in the enterprise will have to be assiduously cultivated. That must start with an open consideration and discussion of the ethics of this work. 10th Anniversary Issue 119

An ethical IDS effort begins with an explicit and transparent governance process. Most public administrative datasets are the legal responsibility of the administering agency so that, typically, these agencies are legally bound to review and approve any request to access their data (Stiles and Boothroyd, 2015). A governance process, usually spelled out in an interagency Memorandum of Understanding (MOU), or some other document, explicitly outlines the process for obtaining agency consent for a given research project, and the minimum criteria (scientific, policy or practice value) that must be met for a project to be considered. The MOU will also usually describe the data security standards that must be adhered to, and the rules and procedures governing access by external entities like researchers. The governance committee might also involve citizen or nongovernment stakeholder participation (i.e., NGO representatives) to assure that the proposed uses of the data are known beyond the internal workings of government or the academy, and meet community standards for value and integrity. Alternatively, an IDS may have a separate community or stakeholder advisory board, with citizen and other stakeholder representation, which provides periodic feedback to the enterprise.

In addition to a governance policy controlling access to data, a policy should be in place requiring that study results be previewed by various stakeholders so that nuances in data and interpretation are discussed. Such a preview and discussion of results contributes to a spirit that the enterprise is jointly motivated by all of the participants in a positive, reform-oriented ethos. If, instead, studies are viewed as padding for academic resumes, or as evaluative ‘gotcha’ efforts to blame agencies for the shortcomings of programmes, agency personnel will be understandably reluctant to participate. Thus, researchers and nongovernmental stakeholders have to respect that, treated insensitively, evaluative results could be perceived as threatening to the careers of public administrators; thus, a clear set of procedures are needed that provide for adequate feedback and buy-in to the results from public agencies.

In general, then, the basic ethical requirement for the operation of an IDS is a communication strategy that is transparent and that engages government data owners along with other community partners with a vested interest in the programmes under study, as well as general citizenry (including sceptics). In this way, people can participate and inform beneficent policies and uses of the data.

As data security protections become more robust, including through data perturbation and remote statistical analysis – approaches that effectively guard against potential re-identification or re-disclosure of personal information, a further ethical consideration is whether government actually has an obligation to contribute data to an IDS or to other important analysis efforts. A government agency may argue that they can’t participate in data sharing or data integration 120 European Journal of Homelessness _ Volume 10, No. 3 _ 2016

efforts because of re-disclosure risks. But these arguments may also be efforts to shield an agency from evaluative assessments, either about a programme’s efficacy or the quality of their data. Again, a shared ethos among participants that efforts are motivated by positive outcomes for citizens, and not for blame, could persuade some reluctant agencies to participate. However, the potential for evasive action does raise the issue of whether or not an assumption should be made (as a value or ethic) that publicly-financed programme data must be made available to these efforts because of the potential of these data for contributing to the common good. In other words, interagency data sharing should be the default position, unless a specific statutory authority can be cited to show that it is prohibited (some prevailing law that restricts some data from any sharing, even for evaluation or audit purposes). Such a value can further contribute to the trans- parency of these efforts, and assure citizens that government cannot hide from uncomfortable truths that may be revealed by their data.

As the use of administrative data for social science and social policy research advances, the ethics of the science of these data must also be considered. First, as with social research in general, research based on administrative data has to meet the human subject protections that generally govern scientific conduct at universities, research institutes and in government through ‘institutional review board’ or ‘ethics committee’ approval. In general, the use of ‘secondary data’ as such represents a much-reduced level of risk to studying participants when compared to primary data collection, with the primary risk being the potential re-disclosure of confidential information to unauthorised parties. Indeed, if appropriate and effective de-identification procedures are employed, such research could be argued as exempt from institutional review board approval in most cases.

However, in the event that administrative data are being used to evaluate interventions, then the interventions themselves should be reviewed by institutional review boards for the ethical nature of the interventions, if not for the use of the administrative data to evaluate them. In many countries, interventions that represent variations in public benefit administration, and other enhancements to benefits as determined by government, may be exempt from institutional review board approval because the primary goal is improved service delivery and not producing generalisable knowledge. In other words, the goal is to evaluate public programmes and their administration, not research to benefit the scientific understanding of a problem. This distinction is important with regard to institutional review board jurisdiction in many communities. However, the potential for government’s understanding of an ‘enhancement’ to be inconsistent with the general public’s understanding does suggest that some level of review should be considered in some cases, or even required, depending on the nature of the intervention. 10th Anniversary Issue 121

A further ethical consideration regarding the use of administrative data in research is that these data were not collected with the intention of supporting research, and they may not be properly suited to that purpose in some cases. An obvious scientific problem is when the data are incorrect or invalid. Data audit routines should look for incomplete variables and out of range codes. More subtly, agency practice may result in variables being coded in ways that do not correspond to their label or original intent. Variables may be used for ‘office’ purposes in ways that might not be readily apparent to researchers, which is another reason for making sure that studies are done collaboratively and results previewed with administrative staff to reduce the likelihood of misinterpretation or misuse of data.

Another subtler, but quite serious ethical consideration regards the re-use of data for purposes beyond their original intent. This is especially sensitive when derivative data analytic products are used to determine eligibility for programmes or are used to render decisions about client access or the receipt of services – decisions that could have quite real and serious implications. Consider a couple of examples. In the US, machine-learning algorithms are currently being used to inform judges’ decisions regarding the sentencing of persons convicted of various crimes. The algorithms are intended to identify the strongest predictors of recidivism given certain sentences, and to remove the bias of judges (Berk and Hyatt, 2015). Variables for race and ethnicity have been intentionally excluded from these algorithms, but former US Attorney General Eric Holder nevertheless expressed concern about these methods (Barry-Jester et al., 2015). According to Holder’s critique, while race may be excluded as an explicit variable, race may nevertheless be embedded in the administrative data because the differential experiences by race of people in the US get effectively encoded in the data through their differential contact with public agencies and programmes. Thus, bias may be ‘hidden’ in the data, which can reinforce bias in sentencing decisions, calling into question the neutrality of the machine-learning algorithms (Crawford, 2016).

In an example closer to the homelessness situation, the use of administrative data to identify ‘high service users’ as a way of prioritising candidates for permanent supportive housing is similarly fraught with potential ethical considerations. First, these data were not collected for this purpose, and may well exclude important information which, were it included, might well lead to a person being determined as eligible as opposed to ineligible for certain housing programmes. For example, the fact that some data may be missing from a given IDS or other system (i.e., data from an outlying county where the services were received) may mean that persons are being determined as ineligible or at least not ‘priority’ when in fact their actual, more inclusive records might indicate otherwise. A second concern relates to whether these ‘high service use’ thresholds actually make sense in determining need, when in fact they reflect service use and not need. This sort of subtle nuance 122 European Journal of Homelessness _ Volume 10, No. 3 _ 2016

is ignored when a simplistic decision rule is derived from crude aggregate utilisation measures. Yet, a third consideration concerns the use of these decision rules to justify leaving some people unserved. The fact that people who are homeless and in need of housing would not receive housing due to some decision tool, like the Vulnerability Index, doesn’t necessarily indicate that the people may be any less eligible under the law to receive that housing. Indeed, it should be no comfort that just because there is a decision rule for rationing limited resources, that a fair result is being meted out. Scarcity may drive rationing, but administrative data shouldn’t be used to justify it or even to decide the winners and losers of rationing, without a clearly legislated (i.e., publicly deliberated) intent and authorisation to do so, along with appropriate protections against inadvertent denial of eligibility due to data limitations.

The use of government-held administrative data, where these contain highly private information and are often accessible only to privileged researchers, calls out for a transparent and ethical framework to guide the conduct of this work. While some of the ethical issues can be anticipated and forthrightly addressed in the policies and procedures of an IDS or a given record linkage project, not all of the issues are as obvious. For that reason, continued engagement with members of the general public, with sceptics (‘devil’s advocates’) and with others knowledgeable about the use and misuse of statistics can help to provide checks and balances for the maximum protection of the data and for the most beneficent uses. 10th Anniversary Issue 123

Conclusions

Homelessness research – and by extension people who experience or are threat- ened with homelessness – could stand to benefit from improved accessibility of administrative data for social policy research. Administrative data offer a potentially low-cost and continuous source of information regarding the prevalence and duration of homelessness in a community over time. Linked with other records, they can allow assessment of the degree to which discharge practices from other social welfare systems are resulting in increased homelessness. Correspondingly, analysts can assess the disproportionate impact of homelessness on other social welfare systems. Interventions intended to reduce rates of homelessness or to expedite stable exits from homelessness can similarly be tracked with administrative records, and the impact on other health and social welfare systems evaluated. These and related efforts offer a new opportunity for the study of homelessness, and for informing changes in policy that could benefit those who are experiencing homelessness or who are otherwise at risk. This new potential for social science and social policy research is not without its own risks, and it requires an explicit and transparent discussion of the ethical considerations of these linked data systems. A strong policy framework and a clear communication and community engagement strategy can provide some significant protections against the potential misuse of these sensitive data. New technological innovations and secure workflow procedures can also be deployed to provide added protections to confidential data and, ultimately, to leverage those data to expand our knowledge of what works best and for whom. 124 European Journal of Homelessness _ Volume 10, No. 3 _ 2016

>> References

Barry-Jester A.M., Casselman, B. and Goldstein, D. (2015) Should Prison Sentences Be Based on Crimes That Haven’t Been Committed Yet? [on-line] Available from: http://fivethirtyeight.com/features/prison-reform-risk-assessment/ [11.07.16].

Benjaminsen, L. and Andrade, S.B. (2015) Testing a Typology of Homelessness across Welfare Regimes: Shelter Use in Denmark and the USA, Housing Studies 30(6) pp.858-876.

Berk, R. and Hyatt, J. (2015) Machine Learning Forecasts of Risk to Inform Sentencing Decisions, Federal Sentencing Reporter 27(4) pp.222-228.

Busch-Geertsema, V., Benjaminsen, L., Filpovic Hrast, M. and Pleace, N. (2014) Extent and Profile of Homelessness in European Member States: A Statistical Update – European Observatory on Homelessness Comparative Studies on Homelessness (Brussels: FEANTSA).

Community Solutions (2016) The 100,000 Homes Vulnerability Index: Prioritizing Homeless People for Housing by Mortality Risk. [on-line] Available from: http://100khomes.org/sites/default/files/About%20the%20Vulnerability%20Index. pdf [17.12.2016].

Crawford, K. (2016) Artificial Intelligence’s White Guy Problem, New York Times, June 25, 2016.

Culhane, D.P., Metraux, S. and Hadley, T. (2002) Public Service Reductions Associated with Placement of Homeless Persons with Severe Mental Illness in Supportive Housing, Housing Policy Debate 13(1) pp.107-163.

Culhane, D.P., Parker, W.D., Poppe, B., Gross K. and Sykes, E. (2008) Accountability, Cost-Effectiveness, and Program Performance: Progress Since 1998, in: D. Dennis, G. Locke and J. Khadduri (Eds.) Proceedings from National Symposium on Homelessness Research (Washington: Department of Health and Human Services and Department of Housing and Urban Development).

Culhane, D.P., Fantuzzo, J., Rouse, H., Tam, V. and Lukens, J. (2010) Connecting the Dots: The Promise of Integrated Data Systems for Policy Analysis and Systems Reform (Philadelphia: Intelligence for Social Policy, University of Pennsylvania).

Fantuzzo, J., Culhane, D.P., Rouse, H. and Henderson, C. (2015) Introduction to the Actionable Intelligence Model, in J. Fantuzzo and D.P. Culhane (Eds.) Actionable Intelligence for Social Policy, pp.1-38. (New York: Palgrave Macmillan). 10th Anniversary Issue 125

Flaming, D., Matsunaga, M. and Burns, P. (2009) Where We Sleep: The Costs of Housing and Homelessness in Los Angeles (Los Angeles: Economic Roundtable).

Government of Canada (2016) National Homelessness Information System. [on-line] Available from: www.esdc.gc.ca/eng/communities/homelessness/nhis/ index.shtml [17.12.2016].

Hamlet, N. (2015) Measuring Health and Homelessness in Fife. [on-line] Available from: www.gov.scot/Resource/0047/00476237.pptx [17.12.2016].

Haskins, R. and Barron, J. (2011) Building the Connection between Policy and Evidence: The Obama Evidence-based Initiatives (London: NESTA).

Hwang, S. (2016) Personal Communication (On file with author).

Jones, K.R., Ford, D.V., Jones, C., Dsilva, R., Thompson, S., Brooks, C.J., Heaven, M.L., Thayer, D.S., McNerney, C.L. and Lyons, R.A. (2014) A Case Study of the Secure Anonymous Information Linkage (SAIL) Gateway: A Privacy- Protecting Remote Access System for Health-related Research and Evaluation, Journal of Biomedical Informatics 50 (August) pp.196-204.

Metraux, S., Byrne, T. and Culhane, D.P. (2010) Institutional Discharges and Subsequent Shelter Use Among Unaccompanied Adults in New York City, Journal of Community Psychology 38(1) pp.28-38.

Nussle, J. and Orszag, P. (2015) Moneyball for Government (New York: Disruption Books).

O’Donoghue-Hynes, B. (2015) Patterns of Homeless Emergency Accommodation Use in Dublin: How do we Compare? Paper presented at European Research Conference, Dublin, 25 September.

Parsell, C., Petersen, M. and Culhane, D.P. (2016) Cost Offsets of Supportive Housing: Evidence for Social Work, British Journal of Social Work doi: 10.1093/ bjsw/bcw115.

Pittini, A., Ghekiere, L., Dijol, J. and Kiss, I. (2015) The State of Housing in the EU 2015 (Brussels: Housing Europe).

Poulin, S., Metraux, S. and Culhane, D.P. (2008) The History and Future of Homeless Management Information Systems, in: R. Hartmann McNamara (Ed.) Homelessness in America Vol. 3, pp.171-180. (Hartford: Praeger).

Rolston, H., Geyer, J. and Locke, G. (2013) Final Report: Evaluation of the Homebase Community Prevention Program (Bethesda, MD: Abt Associates). 126 European Journal of Homelessness _ Volume 10, No. 3 _ 2016

Shinn, M., Greer, A.L., Bainbridge, J., Kwon, J. and Zuiderveen, S. (2013) Efficient Targeting of Homelessness Prevention Services for Families, American Journal of Public Health 103(S2) pp.S324-S330.

Stergiopoulis, V., Gozdzik, A., Misir, A. et al. (2015) Effectiveness of Housing First with Intensive Case Management in an Ethnically Diverse Sample of Homeless Adults with Mental Illness: A Randomized Controlled Trial, PloS ONE 10(7): e0130281. Doi: 10.1371/journal.pone.0130281.

Stiles, P.G. and Boothroyd, R.A. (2015) Ethical Use of Administrative Data for Research Purposes, in: J. Fantuzzo and D.P. Culhane (Eds.) Actionable Intelligence for Social Policy, pp.125-155. (New York: Palgrave Macmillan).

Welsh Government (2015) Supporting People Data Linking Feasibility Study: Emerging Findings Report – Research Summary, No. 65/2105 (Cardiff: Welsh Government). 10th Anniversary Issue 127

Homelessness in Australia: Service Reform and Research in the 21st Century Nicolas Herault and Guy Johnson

The Melbourne Institute for Applied Economic and Social Analysis, University of Melbourne, Australia

Centre for Applied Social Research, RMIT University, Melbourne, Australia

>> Abstract_ In 2008, the newly-elected Australian Government made homelessness its highest policy priority. This resulted in a substantial injection of funding to support new homelessness initiatives designed to prevent homelessness, as well as end chronic homelessness. In addition, over $11m was invested in research. In this paper, we argue that despite some promising service innovations, service reform failed, by and large, due to resistance from entrenched interests committed to maintaining a systemic focus on transitional support tied to short- and medium-term accommodation, and a lack of leadership. Further, we question whether the goal of ending homelessness can actually be met through systems reform only. We then examine the Government’s investment in research. We examine ‘Journeys Home’, a longitudinal study that tracked a sample of vulnerable and homeless households, and identify what makes it so unique. We draw attention to the paradox that despite Journeys Home being recognised around the world as an exceptional dataset capable of answering fundamental questions about the dynamics of homelessness, it has yet to have a meaningful impact on Australian homelessness policy and service directions. We explain why this is so and what needs to be done to harness its potential to drive evidence-informed policy and practice change.

>> Keywords_ Journeys Home, longitudinal research, impact of research on policy and services

ISSN 2030-2762 / ISSN 2030-3106 online 128 European Journal of Homelessness _ Volume 10, No. 3 _ 2016

Introduction

Australia is a prosperous country. Most Australians enjoy a relatively high standard of living: unemployment is relatively low; life expectancy is high and over two thirds of Australians own or are purchasing their own home. Nonetheless, like many other countries, Australia has a problem with homelessness. According to the Australian Bureau of Statistics, over 100,000 Australians were homeless on census night in 2011, and over one in 10 have been homeless at some point in their lives (Australian Bureau of Statistics, 2010; 2012). For most of the last three decades, homelessness programmes and research in Australia have lagged behind developments in the rest of the world. In 2007 the situation started to change. This paper examines what changed, why, and the legacy of that change.

Background

Australian policy interest in homelessness started in the early 1970s with the passage of the Homeless Persons Assistance Act (1974). Under the terms of the Act, NGOs were required to assist chronically homeless persons, most of whom were older men living in inner city areas (DeHoog, 1972; Jordon, 1994). In the early 1980s, the demographic characteristics and geographic distribution of the homeless population started to change (Sackville, 1976). These changes foreshad- owed what some researchers around the world would term the ‘new homeless’ (Hopper, 1997; Huth, 1997; Lee et al., 2010).

The Federal Government subsequently launched the Supported Accommodation Assistance Program (SAAP) in 1985. Services provided under SAAP were not targeted to the ‘skid row’ population, but to young people, families and women, as well as single people. The shift in policy focus also reflected a shift in how homelessness was understood. In the past homelessness was thought to be a chronic condition. However, SAAP was based on the view that homelessness was a temporary crisis that could be addressed through the provision of transitional support linked to short- and medium-term emergency accommodation (Neil and Fopp, 1992). Over the next two decades SAAP underwent a series of formal reviews. And, while these reviews resulted in some modifications to SAAP’s priorities and focus, the principles that guided SAAP, and the underlying conception of homelessness as a temporary crisis, remained the same.

By the middle of the first decade of the 21st century there were around 1,500 SAAP services across Australia receiving approximately AUS$400m in recurrent funding. However, the structural context in which SAAP services operated had changed considerably since it began. The housing market of the early 21st century was very different from the housing market in the 1980s. Across the country, housing 10th Anniversary Issue 129

affordability had become a major issue, and affordable private rental accommodation was increasingly scarce, particularly in major cities. From the mid-1990s funding for public housing declined in real terms, and public housing stock declined both in real terms and as a percentage of the nation’s housing stock1 (Groenhart and Burke, 2014). Although unemployment was relatively low throughout the 1990s, it remained persistently high among some groups, such as young people. The 1990s was also a period characterised by ongoing welfare reforms as the Federal Government applied more stringent targeting measures and increased the obligations and responsibilities of welfare recipients.

Despite clear signals that SAAP agencies were struggling to meet the increasing demand brought about by structural changes (Johnson, 2012), from the mid-1990s policy interest in homelessness was low. What interest there was in homelessness focused on bureaucratic concerns such as service duplication and the lack of service integration rather than any substantive rethinking about the way SAAP services were delivered or funded (Department of Family and Community Services, 2001; Department of Human Services (Vic), 2002). There was no reason to think this situation was going to change. But in late 2007 it did and homelessness suddenly shifted from comparative obscurity to national prominence.

Reform and Research

On 3 December 2007, a new national labour government swept to power after a decade of conservative rule. Within a fortnight, the new Prime Minister declared that homelessness was a ‘national disgrace’ and immediately identified homelessness as the Government’s highest social policy priority. Although it is not entirely clear why homelessness figured so prominently in the new Government’s policy agenda, Parsell and Jones (2014, p.433) mount a strong argument that the focus on homelessness was one way of differentiating the “new government from the old”. Homelessness was framed as a social problem that was inconsistent with Australia’s “espoused egalitarian ethos” (p.428). And, by locating homelessness in a moral framework, the new Government was able to attack the previous government directly for “allowing this morally unacceptable problem to occur” (p.433).2

In a remarkably short amount of time the new Government appointed the first ever Minister for Housing and Homelessness, established a Prime Minister’s Council on Homelessness and, along with state governments, signed off on the National

1 In 1996, public housing accounted for 5.2% of the nation’s housing stock. By 2011 it had declined to 4.1%. 2 Another school of thought is that, as a child, the Prime Minister had been homeless and was deeply affected by his experience. 130 European Journal of Homelessness _ Volume 10, No. 3 _ 2016

Partnership Agreement on Homelessness (NPAH) in November 2008. Under the NPAH, the Council of Australian Governments committed over AUS$800m of additional funding for support services and new homelessness initiatives over four years (2009-2012).3 This amounted to a 55% increase in funding (FaHCSIA, 2008). Crucially, though, these developments were occurring in the wake of the Global Financial Crisis (GFC). To ameliorate the impact of the GFC, the Federal Government initiated the Nation Building: Economic Stimulus Plan. The programme was partly driven by concerns about the impact of the GFC on the housing sector and construction industry, both of which are vital to the nation’s economic wellbeing. As a result of the stimulus programme, nearly 20,000 new social housing dwellings were constructed, many of which were targeted to at-risk and homeless persons.

Underpinning the NPAH was the Federal Government’s White Paper on Homelessness called ‘The Road Home’ (FaHCSIA, 2008). The Road Home committed the Federal Government to two ‘2020’ goals: (i) halving overall homelessness; and (ii) offering supported accommodation to all rough sleepers who need it. The Road Home identified three policy approaches to achieve the two ‘2020’ goals and to measurably reduce the number of people exiting institutional settings and private and social housing into homelessness: early intervention and prevention; better service integration and improved service capacity; and assisting people to sustain their housing. The Road Home also identified the need for innovative, evidence-based services. Indeed, the notion of ‘evidence-based policy’ became a government mantra. Addressing the Fifth National Homelessness Conference, the Prime Minister outlined a vision to “draw out bold new ideas and to identify evidence-based approaches to reduce homelessness” (Kevin Rudd, speech to National Homelessness Conference, 22 May 2008).

Research evidence was positioned in policy discourse as a key link between reducing homelessness and the selection of new homelessness initiatives. This was most evident in relation to the Government’s goal of reducing the numbers of rough sleepers, which it equated with chronic homelessness (Parsell, 2014). Seizing on research evidence from the US of the effectiveness of Housing First and Permanent Supportive Housing (PSH) approaches, the Government committed to funding ‘Street to Home’ services (i.e., Housing First), as well as ‘Common Ground’ facilities (i.e., permanent supportive housing – PSH), in every state and territory. A Housing First/PSH approach to working with chronically homeless persons represented a radical departure from what was possible under SAAP. The Street to Home initiative, which drew on many of the principles articulated by the Pathways to Housing model, had a strong evidence base (Stefancic and Tsemberis, 2007; Tsemberis,

3 The original NPAH has been renegotiated a number of times. There remains some uncertainty as to whether it will continue past its current iteration, due to expire in 2017. 10th Anniversary Issue 131

2011). And, while advocates in Australia and elsewhere arguably oversold aspects of Housing First (Kertesz et al., 2009; Pleace, 2011; Johnson et al., 2012), the selection of ‘Street to Home’ confirmed the Government’s willingness to trial new models, as well as its commitment to policy that drew on sound research evidence. The Government also invested in the ‘Common Ground’ model of permanent supportive housing from the US. The selection of Common Ground further confirmed the Government’s willingness to try new approaches. However, as Parsell et al. (2014) note, the selection of Common Ground was not based on research evidence. Rather “intuition and direct personal experiences were afforded more credibility…. by relevant stakeholders than peer-reviewed research” (p.84). Indeed, the selection of the Common Ground model serves as a timely reminder that research evidence is not always at the top of policy-makers’ ‘knowledge hierarchy’ (Parsell et al., 2014).

The Road Home nonetheless reported that the development and delivery of effective service responses was hampered by gaps in the existing evidence base, most notably a lack of larger-scale longitudinal data. Although Australian researchers had made substantial progress through the 1990s and early 21st century, studies were primarily qualitative and very small. Where quantitative studies had been undertaken they, too, were small, cross-sectional and restricted to specific groups of currently homeless persons or services users.

Thus, alongside service innovation a key element in the Government’s strategy to reduce homelessness was the development of a national research agenda. After 12 months of consultation with researchers, service providers and policy-makers, the Federal Government released The National Homelessness Research Agenda 2009-2013 in November 2009 with the aim “[t]o improve the evidence base for preventing and responding to homelessness” (FaCHSIA 2009, p.4). The research agenda identified several core priorities for future research: to inform and improve the service system and practice; to increase understanding of homelessness; and to improve data and the measurement of homelessness.

Previous governments had also noted a lack of research evidence, but as they hadn’t done anything about it, there was considerable scepticism as to whether this government would be any different. This changed when the Federal Government announced it was allocating AUS$11.4m to The National Research Agenda – by far the largest single investment in homelessness research in Australia. The Government used the funding to support three separate initiatives. First, it allocated AUS$1.5m to 16 small research projects that were of national significance and focused on priorities identified in the research agenda. The 16 projects examined a broad range 132 European Journal of Homelessness _ Volume 10, No. 3 _ 2016

of topics, covering areas such as sole fathers, families and children, service integration, institutional costs of homelessness, unemployment and the clinical care needs of chronically homeless persons.

Second, the Government allocated over AUS$4m to three research partnerships. The aim of these partnerships was to deliver an agreed programme of research that focused on more complex multi-year projects. Flinders University, the University of Queensland and Swinburne University were awarded the contracts. Research produced from these partnerships included longitudinal evaluations of recently funded Housing First and early intervention initiatives, as well as studies that examined youth, later life and indigenous homelessness, as well as homelessness prevention for women and children.

Together, the two research initiatives contributed new evidence about homeless subgroups, service capacity, and the effectiveness or otherwise of specific homelessness interventions. This was useful evidence for policy-makers and providers. However, the scope of these projects was limited and none could provide any robust information on the factors that contribute to the onset of homelessness, on whether conditions related to the onset of homelessness are also associated with its persistence, or on the factors that contribute to exits from homelessness.

Australia was not alone in being unable to provide reliable answers to these questions. Clear answers about the causes and consequences of homelessness have largely eluded researchers around the world due to a lack of appropriate data. A small number of researchers from the US and Denmark have successfully used administrative and longitudinal studies to shed light on various aspects of the dynamics of homelessness (Kuhn and Culhane, 1998; Metraux et al., 2001; Benjaminsen and Andrade, 2015; Benjaminsen, 2016). But these studies have a number of weaknesses that limit their capacity to illuminate the determinants of entries and exits from homelessness. Most notably, the information that administrative datasets capture is often limited and the quality of data uneven, and they only provide information on people who use particular services.

The Government’s third initiative aimed to address this gap. Just over AUS$5m was subsequently awarded to The Melbourne Institute of Applied Economic and Social Research at The University of Melbourne to undertake a large-scale, national longitudinal study that focused on housing instability and homelessness, subsequently called ‘Journeys Home’. The commissioning of Journeys Home was a major step forward. For the first time researchers were going to have sufficient funding to attempt what had never been done before – a longitudinal survey that tracked a national sample of individuals exposed to high levels of housing insecurity and that employed more rigorous sampling methods than previously used. 10th Anniversary Issue 133

Journeys Home: What Makes it Unique?

Commissioning Journeys Home was a bold move. Although the financial investment was bigger than ever seen before, three significant challenges had to be addressed if questions about causes and consequences were to be successfully answered.

First, the dataset had to include housed people and those at risk as well as homeless households. The inclusion of vulnerable households is important because it provides the opportunity to identify the factors that precipitate the onset of homelessness. Crucially, it also gives researchers a control group to compare homeless people against, given that such a specific population can hardly be deemed comparable to the general population covered in other household surveys. However, drawing a sample of individuals vulnerable to homelessness is difficult because homelessness is such a rare occurrence. This means that in any sample randomly drawn from a vulnerable population, the likelihood that an individual will experience homelessness is very low. Journeys Home was faced with a similar problem when it drew a sample from the Research Evaluation Database (RED), which contains administrative records for all Centrelink4 income support recipients. Centrelink provides all income support payments to eligible members of the Australian community, and most people at risk of or experiencing homelessness would likely be in receipt of a Centrelink payment. The problem is that nearly five million Australians receive Centrelink payments at any given time, and the majority of these people would not be at risk of homelessness. Fortunately, there was a way around this problem.

Since 2010, Centrelink staff have been required to flag in their database customers they assess to be either ‘homeless’ or ‘at risk of homelessness’.5 This provided the opportunity to draw a sample of people who were homeless, had recently experienced homelessness, or were at risk of homelessness. All Centrelink customers aged 15 and over in receipt of benefits during a 28-day period prior to 27 May 2011 were considered to be in the in-scope population. This population contained 27,017 individuals flagged as homeless and 15,319 individuals flagged as at risk of homelessness.

A limitation of the Centrelink flags is that the flagging protocols were not consistently followed by Centrelink staff across the country. Consequently, a decision was made to identify a third group that had characteristics similar to those identi-

4 Centrelink delivers a range of government payments and services for retirees, the unemployed, families, carers, parents, people with disabilities, Indigenous Australians, and people from diverse cultural and linguistic backgrounds, and provides services at times of major change. The majority of Centrelink’s services are the disbursement of social security payments. 5 The definitions of ‘homeless’ and ‘at risk’ used by Centrelink broadly matches the definition put forward by Chamberlain and MacKenzie (1992). 134 European Journal of Homelessness _ Volume 10, No. 3 _ 2016

fied by Centrelink as ‘homeless’ or ’at risk’ but that had not been flagged. Using a logistic regression equation, 95,755 persons were identified who were, at least in a statistical sense, vulnerable to homelessness. The extensive list of predictor variables was largely driven by what was available in the RED data and included controls for key demographic characteristics – health, housing tenure type, income and income support history, prior incarceration and a range of other indicators used by Centrelink to identify ‘vulnerability’, such as drug or alcohol dependence, a lack of literacy and language skills and having experienced a recent traumatic relationship breakdown.

Thus, the total population from which Journeys Home drew its final sample was 138,181 individuals. Individuals were randomly selected from each of the three subgroups. After determining various individuals and groups to be ‘out of scope’, 6 2,719 individuals were randomly selected. Almost 62% of this group (n=1,682) agreed to participate. Individuals classified as homeless, at risk and vulnerable accounted for 35%, 37% and 28% of the sample, respectively (for further details on the population and sampling methodology see Wooden et al., 2012). The response rate to the initial survey (wave 1) not only compares favourably with other studies that sample from seriously disadvantaged populations (O’Callaghan et al., 1996; Randall and Brown, 1996; Weitzman et al, 1990), but it is also in line with panel surveys of the general population, including the Household Income and Labour Dynamics in Australia (HILDA) survey, 7 the German Socio-economic Panel Study, and the British Understanding Society Survey, which have wave 1 response rates of 61%, 66% and 57%, respectively (Watson and Wooden, 2014).

The second challenge was to draw a sample from multiple locations. To understand, for instance, how the labour or housing market affects entries and exits from homelessness, one requires variations in the conditions of the housing and labour markets. The more the better. Studies that sample from a single location, or even a small number of locations, are unable to exploit variation in local conditions and hence are unable to estimate to what extent they affect homelessness. Journeys Home drew its wave 1 sample from 36 different locations (or clusters) covering city, suburban, regional and remote areas from all states and territories

6 Those out-of-scope where those identified as: (i) in prison; (ii) an overseas customer; (iii) requiring an interpreter; (iv) having specifically indicated to Centrelink that they were not willing to participate in research studies; or (v) having a record marked as ‘sensitive’. 7 The Household, Income and Labour Dynamics in Australia Survey (HILDA) is a household-based panel survey. It began in 2001 and 14 waves have been completed. Surveys are conducted annually. Funding is guaranteed for 18 waves. HILDA collects information about economic and subjective well-being, labour market dynamics and family dynamics. The wave 1 panel consisted of 7,682 households and 19,914 individuals. In wave 11 this was topped up with an additional 2,153 households and 5,477 individuals. 10th Anniversary Issue 135

across the country. This means that for the first time, a national sample of those who were homeless or at risk of homelessness was constructed. National coverage was maintained in subsequent waves by tracking wave 1 respondents, even if they moved out of the 36 original clusters. Journeys Home went beyond drawing a sample from multiple locations in that it offered national coverage, not unlike general household panel surveys such as the HILDA (Household, Income and Labour Dynamics in Australia) survey. National coverage allows researchers not only to exploit variation in local conditions with respect to the housing market, the labour market or even state and territory public policies, it also allows researchers to draw inferences at the national level. This represents a major step forward in addressing the limited geographical applicability of previous studies. For the first time, it is possible to examine the determinants of homelessness not only in a few urban areas but in an entire country, thanks to a survey that consistently measured how entries into homelessness, the experience of homelessness, and the characteristics of the homeless population vary across urban and rural areas in Australia.

The third and last major challenge was to generate a high-quality longitudinal dataset. Researchers have long been aware that the best way to answer questions about entries and exits from homelessness is through longitudinal research. However, maintaining contact with vulnerable and homeless individuals can be challenging given the relatively high rates of mobility, mortality and imprisonment in this population. Attrition is problematic for two reasons. The first obvious reason is that it reduces sample size. Large attrition rates can preclude any longitudinal analysis because there are simply not enough individuals that can be followed through time. More importantly, attrition raises particular concerns when it is not random. Non-random attrition occurs when individuals dropping out of the sample are different in observable or unobservable ways from those staying in the sample. And non-random attrition is particularly concerning when attrition rates are high. The evaluation of dynamic patterns, whether they relate to homelessness, health, income, geographical mobility or any other changing characteristics, can be biased if the underlying survey exhibits substantial attrition that is not random.

Journeys Home was extremely successful in retaining the original participants. Journeys Home conducted six surveys, with interviews held every six months. Over the 2.5-year observation period, 91% (wave 2), 88% (wave 3), 86% (wave 4), 85% (wave 5) and 84% (wave 6) of the original wave 1 respondents were re-interviewed. The very high retention rate means that sample size is unlikely to be an issue for any longitudinal analysis and it gives confidence that findings based on Journeys Home suffer little from non-random attrition biases. These high retention rates are most likely due to two factors: all sample members were offered a AUS$40 incentive 136 European Journal of Homelessness _ Volume 10, No. 3 _ 2016

each time they agreed to be interviewed and the organisation sub-contracted to undertake the field work had access to participants’ most recent contact details, including their address and phone number, from the Centrelink database.

The longitudinal dimension also allows researchers to control for both observed and unobserved individual heterogeneity in multivariate analyses. Econometric models such as fixed and random effects models that account for unobserved heterogeneity require longitudinal data. And, although Journeys Home includes a wide range of individual and household characteristics, it cannot possibly cover the full spectrum of characteristics relevant for any outcome variable. For instance, there are unobserved characteristics that predispose some people to homelessness or to longer durations of homeless, and failure to account for this unobserved heterogeneity could very well lead researchers to the wrong conclusions.

While these three features alone distinguish Journeys Home from other datasets available to homelessness researchers, a number of features of the survey design are also worth mentioning. The survey tool(s) used in Journeys Home were designed to elicit information in a number of areas that were thought to have a bearing on entries and exits, but which are not commonly captured in administrative data. Hence, questions about social network composition, employment, service contact, health and well-being, family history, exposure to violence, as well as housing and living arrangements provide researchers with an extremely rich dataset.

The survey was also developed in such a way that it avoided any specific definition of homelessness; indeed, the term ‘homelessness’ is never used. Rather, the instrument collected a raft of information on where people were staying, as well as the stability and the quality of their accommodation. The benefit of this is that researchers can apply their own definition of homelessness, as well as test different definitions.

Another distinctive feature of the Journeys Home survey tool is that it included a housing calendar designed to capture all changes in housing circumstances between interviews. Beginning in wave 2, Journeys Home respondents who had a new address (or had lived in other places) since their last interview were asked to report the month and period (beginning, middle or end of month) that they had left their last accommodation and the type of accommodation they had moved into. This line of questioning was repeated to capture all subsequent moves prior to the respondent moving into their current accommodation at the time of the interview. When individuals were sleeping rough, the survey only captured the timing of their move into accommodation rather than their moves from one street to another. These data provide very detailed information about the timing of homeless and housing spells, which is crucial in determining any causal relationships. 10th Anniversary Issue 137

Finally, drawing the sample from Centrelink provided the opportunity to link respondents’ survey responses to their administrative records. The administrative records contain detailed and accurate information on the benefits history, accommodation types and any periods of incarceration of respondents, as well as any medical conditions while receiving benefit payments. A question seeking consent to link respondents’ survey responses to their Centrelink records was included and 98% of respondents consented. This is a key feature of Journeys Home because it allowed accurate information to be collected from administrative sources on issues that some respondents may be reluctant to reveal, such as past incarceration or medical conditions. Benefits history are also fundamental to determine present and past benefit payments, as recall bias, the complexity of the transfer system, and the potential stigma associated with these payments often translate into very poor measures of these payments in surveys relying on self-reported information.

Taken together, Journeys Home provides researchers with the opportunity to rigor- ously interrogate many important questions about the dynamics of homelessness using more sophisticated techniques than previously. For instance, Cobb Clark and colleagues (2016) utilise the full six waves and apply survival analysis to model exits from homelessness using two different definitions of homelessness – a literal definition and a broader definition, which they term ‘housing insecurity’. Importantly, they apply techniques that account for time invariant and unobserved heterogeneity, which is a unique contribution. And it makes a difference. As with other studies, they find that exit rates exhibit negative duration dependence (i.e., exit rates decrease with the length of time spent homeless) when individual specific heterogeneity is ignored. However, when they control for unobserved and observed heterogeneity, such as demographic characteristics, education, health and a range of other background characteristics, they find “evidence of significant positive duration dependence in the initial stages [… ], with exit rates then falling for longer durations” (Cobb-Clark et al., 2016, p.66). Hence, they conclude that “the common wisdom that exit rates fall continuously with increased spell length due to a combination of selectivity and scarring effects appears to be overly simplistic” (p.67).

Another important contribution comes from McVicar et al. (2015), who investigated the relationship between substance use and homelessness. Using fixed effects modelling to examine four waves of Journeys Home data, they found homelessness and substance misuse to be closely related. The relationship was, however, driven “by observed and unobserved individual characteristics which cause individuals to be both more likely to be homeless and to be substance users” (p.89). Other longitudinal studies could not control for observed and unobserved individual characteristics, but Journeys Home can. And it makes a difference, as it led the authors to the conclusion that the association between homelessness and substance use is unlikely to be causal in either direction. Indeed, these and other papers (Scutella 138 European Journal of Homelessness _ Volume 10, No. 3 _ 2016

et al., 2013; Johnson et al., 2015), have made an important contribution to the evidence base in Australia, and perhaps internationally. With a raft of manuscripts currently in progress covering topics including health, psychological distress and crime, the contribution of Journeys Home is set to increase.

Lessons

2008 was a watershed year for homelessness in Australia. And now, eight years later, the legacy is clear to see. Service reform was overdue, but the opportunity was, by and large, wasted. According to the Australian Bureau of Statistics, the number of people experiencing homelessness increased by 17% between the 2006 and 2011 censuses (Australian Bureau of Statistics, 2012), and the number of rough sleepers has also increased, with rough sleeping now much more visible in many major cities across the country (City of Melbourne, 2016). The innovative and often evidence-based ideas that underpinned reform in the area of chronic homelessness have not taken hold in the way that was hoped. In part, this is because supporters of Housing First and PSP failed to bring along existing service providers. Indeed, by critiquing the existing system without recognising the historical, material and structural constraints that faced existing service providers, those services were alienated from the broader process of change. However, it is equally true that existing providers were overly defensive at times and many sought to preserve the status quo. Despite the promise of change, Australia’s primary response to homelessness remains much the same – the majority of funding is still directed towards transitional support linked to short- and medium-term emergency accommodation. Another reason system reform stalled was because of a lack of leadership in key areas. Federal and state advisory councils were populated and often chaired by prominent services providers who conflated their own interests with the needs of the broader system. This created much rancour and resistance at the time. A key lesson is that it is important to have independent non-aligned people and institutions driving policy and practice change. The bigger lesson is that without structural reform increasing the supply of affordable housing, the capacity of systems reform to reduce homelessness, let alone end it, is limited (Bullen and Reynolds, 2014).

The story with research is slightly different and there are many positive lessons to learn. Governments often spend small sums of money on minor research projects that are not particularly rigorous or reliable. Of all the studies funded under The National Research Agenda, Journeys Home has the most potential to provide policy-makers with the evidence they need to develop and deliver effective service responses. 10th Anniversary Issue 139

While the greatest value of Journeys Home lies in having a large robust panel dataset capable of producing robust, policy-relevant evidence, Journeys Home has also engaged an entirely new set of researchers. Whereas in the past homelessness research in Australia was dominated by a small group of researchers from limited disciplinary backgrounds, Journeys Home has introduced homelessness to new disciplines such as clinical psychology, economics, econometrics, criminology, quantitative sociology and social work. Not only do new researchers bring fresh ideas, they bring methodological skills previously unavailable.

Another sensible decision was aligning some Journeys Home questions with other surveys. Journeys Home borrowed many questions from HILDA. Not only was there a benefit in that some questions had already been tested, but using a selection of HILDA questions provides an opportunity to make comparisons between vulnerable and mainstream populations.

Another important decision was to make Journeys Home a publically available dataset. In doing so, Journeys Home has connected Australian researchers with some of the best researchers in the world. Engaging with overseas researchers has opened many new and interesting lines of enquiry, both empirical and theoretical. For instance, Dan O’Flaherty from Columbia University in the US, along with Yi-Ping Tseng (The University of Melbourne) and Rosanna Scutella (RMIT University), are using Journeys Home to examine whether private information held by people better predicts homeless entries than public information agencies can obtain. The findings have potentially important implications for service delivery, particularly in the area of assessment, which remains a problematic policy and practice issue in Australia.8

While the Journeys Home research team got many things right, any future attempt to undertake a similar study here or overseas might consider a few changes. The two-and-a-half-year observation period is arguably too short – doubling it to five years, while costly and increasing the risk of attrition, would undoubtedly be better. The addition of a fourth group drawn from the general Centrelink population and not at any significant risk of homelessness would strengthen the study.9 In addition, there should be more consideration in the planning phase about possible links to national administrative datasets. In Australia, linking Journeys Home to Medicare (health) and homeless service systems data (SHIP) would have provided immensely valuable insights into patterns of service use among vulnerable and homeless households, and the associated costs.

8 Paper presented at 2016 Workshop on Homelessness and Housing Insecurity. University of Melbourne, 18-19 July 2016. 9 Suggested by two of the key Journeys Home researchers: Rosanna Scutella and Yi-Ping Tseng. 140 European Journal of Homelessness _ Volume 10, No. 3 _ 2016

Journeys Home has started to make an important contribution to the study of the dynamics of homelessness. Australian policy-makers are now getting access to robust findings identifying key risk factors for homelessness and factors that prolong homelessness, as well as findings on the impact of structural factors such as housing and labour markets. Importantly, Australian policy-makers and service providers no longer have to rely on studies from other countries where social, economic and cultural conditions differ. Indeed, Journeys Home provides the sort of nuanced localised findings that are crucial to developing and delivering services responses that meet the needs of vulnerable Australians.

However, it is clear that research from Journeys Home has yet to filter down and inform policy and practice decisions. While it is still early days, much more work needs to be done by academic researchers to disseminate their findings to non- academic audiences. While the scholarly potential of Journeys Home is rich, the true strength of Journeys Home lies in its potential to drive evidence-informed policy and practice change. Harnessing this potential should be a priority for researchers, policy-makers and service providers.

Journeys Home is by no means perfect, but it is a good example of what can happen when governments and researchers collaborate in the truest sense of the word. Journeys Home has put Australian research firmly in the spotlight across the world. And perhaps this could be its most important legacy: Journeys Home demonstrates that Australia does not always have to follow the rest of the world, but can occasionally lead it as well. 10th Anniversary Issue 141

>> References

Australian Bureau of Statistics (2010) General Social Survey: Summary Results, Australia, 2010 (Canberra: Australian Bureau of Statistics).

Australian Bureau of Statistics (2012) Census of Population and Housing: Estimating Homelessness (Canberra: Australian Bureau of Statistics).

Benjaminsen, L. (2016) Homelessness in a Scandinavian Welfare State: The Risk of Shelter Use in the Danish Adult Population, Urban Studies 53(10) pp.2041-2063.

Benjaminsen, L. and Andrade, S. (2015) Testing a Typology of Homelessness across Welfare Regimes: Shelter Use in Denmark and the US, Housing Studies 6(30) pp.858-876.

Bullen, J. and F. Reynolds (2014) Reforming the Service System, in: C. Chamberlain, C., Johnson, G. and C. Robinson (Eds.) Homelessness in Australia: An Introduction pp.273-292. (Sydney, UNSW Press).

Chamberlain, C. and Mackenzie, D. (1992) Understanding Contemporary Homelessness: Issues of Definition and Meaning, Australian Journal of Social Issues 27(4) pp.274-297.

City of Melbourne (2016) StreetCount 2016: Final Report (Melbourne City of Melbourne).

Cobb-Clark, D., Herault, N., Scutella, R. and Tseng, Y. (2016) A Journey Home: What drives how Long People are Homeless?, Journal of Urban Economics 91 pp.57-72.

Dehoog, J. (1972) Skid Row Dossier (Melbourne: Sun Books).

Department of Family and Community Services (2001) National Homelessness Strategy (Canberra: Department of Family and Community Services).

Department of Human Services (Vic) (2002) Victorian Homelessness Strategy: Action Plan and Strategic Framework. Directions for Change: A Collaborative Approach to Improving our Response to Homelessness (Melbourne: Department of Human Services).

FaHCSIA (2008) The Road Home: A National Approach to Reducing Homelessness (Canberra: Department of Families, Housing, Community Services and Indigenous Affairs). 142 European Journal of Homelessness _ Volume 10, No. 3 _ 2016

FaHCSIA (2009). National Homelessness Research Agenda 2009-2013. (Canberra: Department of Families, Housing, Community Services and Indigenous Affairs).

Groenhart, L. and Burke, T. (2014) Thirty Years of Public Housing Supply and Consumption 1981 – 2011 (Melbourne: Australian Housing and Urban Research Institute).

Hopper, K. (1997) Homelessness Old and New: The Matter of Definition, in: D. Culhane and S. Hornburg (Eds.) Understanding Homelessness: New Policy and Research Perspectives pp.9-67. (Washington, D.C: Fannie Mae Foundation).

Huth, M. (1997) America’s New Homeless, in: M. Huth and T. Wright (Eds.) International Critical Perspectives on Homelessness pp.41-54 (Westport, Connecticut: Praeger).

Johnson, G. (2012) Housing First ‘Down Under’: Revolution, Realignment or Rhetoric?, European Journal of Homelessness 6(12) pp.183-191.

Johnson, G., Parkinson, S. and Parsell, C. (2012) Policy Shift or Program Drift: Implementing Housing First in Australia (Melbourne: Australian Institute of Housing and Urban Research).

Johnson, G., Scutella, R., Tseng, Y. and Wood, G. (2015) Entries and Exits from Homelessness: A Dynamic Analysis of the Relationship between Structural Conditions and Individual Characteristics (Melbourne: Australian Housing and Urban Research Institute).

Jordon, A. (1994) Going Bad: Homeless Men in an Australian City (Melbourne: The Council to Homeless Persons).

Kertesz, S., Crouch, K., Milby, J., Cusimano, R. and Schumacher, J. (2009) Housing First for Homeless Persons with Active Addiction: Are we Overreaching?, The Milbank Quarterly 87(2) pp.495-534.

Kuhn, R. and Culhane, D. (1998) Patterns and Determinants of Public Shelter Utilization among Homeless Adults in New York City and Philadelphia, Journal of Policy Analysis and Management 17(1) pp.23-44.

Lee, B., Tyler, K. and Wright, J.D. (2010) The New Homelessness Revisited, Annual Review of Sociology 36 pp.501-521.

McVicar, D., Moschion, J. and Van Ours, J. (2015) From Substance Use to Homelessness or Vice Versa, Social Science and Medicine 136-137 pp.89-98. 10th Anniversary Issue 143

Metraux, S., Culhane, D., Raphael, S., White, M., Pearson, C., Hirsch, E., Ferrell, P., Rice, S., Ritter, B. and Cleghorn, S. (2001) Assessing Homeless Population Size through the use of Emergency and Transitional Shelter Services in 1998: Results from the Analysis of Administrative Data from Nine US Jurisdictions, Public Health Reports 116(4) pp.334-352.

Neil, C. and Fopp, R. (1992) Homelessness in Australia: Causes and Consequences (Melbourne: CSIRO).

O’Callaghan, B. and Dominian, L. with Evans, A., Dix, J., Smith, R., Williams, P. and Zimmeck, M. (1996) Study of Homeless Applicants (London: Department of the Environment).

Parsell, C. (2014) Chronic Homelessness: A Political and Moral Priority, in: C. Chamberlain, G. Johnson and C. Robinson (Eds.) Homelessness in Australia: An Introduction, pp.233-252. (Sydney, NSW: UNSW Press).

Parsell, C. and Jones, A. (2014) Bold Reforms or Policy Overreach? : Australia’s Attack on Homelessness, International Journal of Housing Policy 14(4) pp.427-443.

Parsell, C., Fitzpatrick, S. and Busch-Geertsema, V. (2014) Common Ground in Australia: An Object Lesson in Evidence Hierarchies and Policy Transfer, Housing Studies 29(1) pp.69 -87.

Pleace, N. (2011) The Ambiguities, Limits and Risks of Housing First from a European Perspective, European Journal of Homelessness 5(2) pp.113-127.

Randall, G. and Brown, S. (1996) From Street to Home: An Evaluation of Phase 2 of the Rough Sleepers Initiative (London: HMSO).

Sackville, R. (1976) Homeless People and the Law: A Report by the Commission for Law and Poverty (Canberra: Australian Government Publishing Service).

Scutella, R., Johnson, G., Mocschion, J., Tseng, Y. and Wooden, M. (2013) Understanding Lifetime Homeless Duration: Investigating Wave 1 Findings from the Journeys Home Project, Australian Journal of Social Issues 48(1) pp.83-110.

Stefancic, A. and Tsemberis, S. (2007) Housing First for Long-term Shelter Dwellers with Psychiatric Disabilities in a Suburban County: A Four-year Study of Housing Access and Retention, Journal of Primary Prevention 28(3-4) pp.265-279.

Tsemberis, S. (2011) Implementing and Understanding the Housing First Approach, Housing First Workshop, Sydney. 144 European Journal of Homelessness _ Volume 10, No. 3 _ 2016

Watson, N. and Wooden, M. (2014) Re-engaging with Survey Non-respondents: Evidence from Three Household Panels, Journal of the Royal Statistical Society A 177(2) pp.499–522.

Weitzman, B., Knickman, J. and Shinn, M. (1990) Pathways to Homelessness among New York City Families, Journal of Social Issues 46(4) pp.125-140.

Wooden, M., Bevitt, A., Chigavazira, A., Greer, N., Johnson, G., Killackey, E., Mocschion, J., Scutella, R., Tseng, Y. and Watson, N. (2012) Introducing Journeys Home, Australian Economic Review 45(3) pp.326-336. 10th Anniversary Issue 145

Homeless Non-Governmental Organisations and the Role of Research Mike Allen

Focus Ireland, Dublin, Ireland

>> Abstract_This paper explores the reciprocal and often complex relationships between homelessness research and NGOs, with a particular focus on the role and motives of homeless NGOs in commissioning and undertaking research and the impact this research has on NGO practice. The paper also explores the influence on NGOs of research emanating from the broader research community, using Focus Ireland, an Irish Homeless NGO, as a case study. The ethical challenges of NGO homelessness research reflect those of the wider research community whether academic or state sponsored, with additional unease regarding the potential exploitation of NGO clients. Citing such concerns, some NGOs refuse to engage in research, arguing that their primary role is to address clients’ personal problems; however, such a stance may unwittingly buttress individual interpretations of homelessness with a consequent downplaying of structural issues. The paper suggests that impact of research findings on the practice of service delivery is rarely direct and sometimes hard to detect. The paper concludes with identifying possible conduits for the communication and transfer of research findings.

>> Keywords_NGO research on homelessness, ethical concerns, communication of research findings

ISSN 2030-2762 / ISSN 2030-3106 online 146 European Journal of Homelessness _ Volume 10, No. 3 _ 2016

Introduction

The European Journal of Homelessness aims to provide “a critical analysis of policy and practice on homelessness in Europe for policy makers, practitioners, researchers and academics.” While a number of articles over the years have explored the relationship between research and policy-making at Government and regional levels, rather less attention has been given to the relationship between research and the organisations where most actual practice takes place – nongovernmental organisations (NGOs) providing services to people who are homeless. It is timely in this reflective edition of the Journal to look at the relationship between these organisations (which I shall refer to, for convenience, as ‘homeless NGOs’) and research carried out into the causes, effects and solutions to homelessness.

There are two broad dimensions to this relationship. The first concerns the extent to which the growing body of research available has impacted on the practice of NGOs: do the findings of research programmes find their way into the practice of NGOs services to the homeless? And if so how does this transfer take place? There is limited documentation on this question, so I will reflect on the extent to which an understanding of the flows into and out of homelessness and the effectiveness of Housing First approaches have been absorbed within the sector.

The second question relates to the role of homelessness NGOs in commissioning and undertaking research. Some homeless NGOs invest a significant portion of their resources in research and NGO-funded research makes a significant contribution to the overall research undertaken into homelessness. However, by no means all homeless NGOs engage in research in this way. So, why do some NGOs invest scarce resources in this way? What are they trying to achieve? How does the role of research fit into the range of other functions carried out by homeless NGOs? Of equal interest, perhaps, is the question of why other homeless organisations do not invest in this area?

In exploring these questions, I will draw on the 30-year history of my own organization Focus Ireland. Over that time, Focus Ireland has had a commitment to commissioning and carrying out its own research, though this has expressed itself in different ways over the period. While I do not claim Focus Ireland’s experience is typical of the approach of other NGOs, I think it serves to demonstrate a range of approaches and challenges which are illustrative throughout the entire sector. The intention overall is to stimulate debate in future issues of the Journal which might throw some additional light on these issues. 10th Anniversary Issue 147

Homeless NGOs

An exploration of why homeless NGOs invest in research needs to start from an understanding of the nature of modern NGOs. Depiction of homeless charities, understandably, tends to concentrate almost exclusively on the ‘front-line services’, the engagement with and accommodation of people who are homeless. However contemporary homeless NGOs are complex organisations with a range of functions – Fundraising, Human Resources, Communications, Marketing and Advocacy as well as Services.

Given the decision-making processes involved, a number of different interests will have to be convinced of the value of any proposed research project before it is approved. The research must fit into some plan – either to modify or launch an internal service or to garner support or funds for the service, or in some cases to seek changes in public policies.

The fact that research is commissioned with some intended purpose can itself undermine the credibility of NGO research. Fitzpatrick and Christian (2006) speak for many when they comment that NGOs ‘can be more concerned with providing effective campaigning material than with obtaining reliable evidence’. Their observation can be interpreted as ‘campaigning intentions result in poor research’ but I don’t see the problem here as an intention to generate campaign material, but rather that the NGOs are seen to have allowed this objective to deflect them from the essential task of research: credibility. In my experience, the more important lesson is that ‘poor research results in poor campaign materials.’

The investment of homeless NGOs in research can be significant – 40% of Irish research into homelessness over the almost 40-year period from the 1970s until 2008 was commissioned by the voluntary sector (O’Sullivan, 2008) – so it matters whether the output is considered to be credible.

There is no reason to single out NGO research for special scepticism. Jacobs et al. (1999, p.11) characterised all research on homelessness as an attempt by ‘vested interests’ to ‘impose their particular definition on policy debates and to push the homelessness issue as they define it either higher up or lower down the policy agenda’. In relation to state funded research, we could easily hypothesise a tendency to minimise the level of homelessness and the extent to which it is caused by state policies rather than the poor choices of the people affected.

Similar reservations can, of course, be applied to ‘academic research’. A quick search on Google Scholar for the term ‘homelessness’ finds 12,300 scholarly articles which were published in 2015 alone. Most of these papers need to be approached with a caution concerning the, sometimes arcane, theoretical framework they explore, or the limitations of their methodology or of their data… and so on. 148 European Journal of Homelessness _ Volume 10, No. 3 _ 2016

The truth is that any commissioner of research is going to have some expectation that it will have some purpose, and this risk of bias (either conscious or not) is the reason for the emphasis on clear and robust methodologies. However, the particular scepticism directed at NGO funded research has led NGOs – who are serious about this element of their work – to adopt a number of techniques aimed at providing confidence about the reliability of research outputs.

Two main tactics are employed to achieve this – external commissioning of independent experts and the creation of research committees comprising individuals who lend their professional reputation to the project. For example, from the beginning Focus Ireland established a ‘Social Policy Committee’ which comprised people with recognised academic credibility, one of whom (Mary McAleese) was later to be elected as President of Ireland. This structure has been continued over the history of the organisation and many of the most respected social policy researchers in Ireland have served on the committee over the years. The Chair of the committee is also a member of the organisation’s Board, bringing this expertise and credibility to the most senior level.

Nevertheless, even if the most rigorous standards are maintained in the actual research, the selection of the aspect of homelessness to be researched and the research question are inevitably framed by the objectives of the organisation commissioning them. It might be useful then to look behind the broad label of ‘providing effective campaigning material’ at the range of possible intentions that NGOs may have in commissioning research, and the effects of this.

Internal and Externally Facing Research

The most common engagement between NGOs and the research community is though commissioned evaluations of services. In general evaluations commissioned by homeless NGOs are internally focused – they look at the services which the NGO delivers to ascertain whether they are effective. Much of what homeless NGOs describe as their ‘research budget’ goes on such activities. The impact of such internally focused evaluations on public policy can be limited and, in the all too many cases where the NGO does not publish the result, non-existent.

In the case of Focus Ireland, there has been a commitment through the life of the organisation to regular evaluation of services, with evaluations published and senior staff given responsibility to ensure that their recommendations are implemented. The growing prevalence of competitive tendering of homeless services can create a reluctance to publish internal evaluations, for fear of revealing sensitive information to competitors. This is unfortunate as high quality internal evaluations can be used by homeless NGOs to gain broader public policy support for innovative 10th Anniversary Issue 149

practice. In my experience, openness to undertaking and publishing critical self- evaluation is linked to a genuine broader interest in conducting and publishing externally facing research. At the very least, an evaluation programme develops expertise in constructing effective research questions and commissioning researchers. The current shift from ‘process evaluation’ to ‘outcome’ and ‘impact’ evaluations, as well as practices such as ‘social return on investment’, mean that such evaluations have the capacity to make a significant contribution to public policy (Royce et al., 2015).

Homeless NGOs have also used the tactic of commissioning evaluations of public policy as a tool to influence how policy develops. Rhodes and Brooke (2010) explore the role of evaluation on homeless policy in Ireland from 1988 until 2008. In the first phase between 1988-95, two of the three evaluations they identify were commissioned by homeless NGOs (Focus Ireland and the National Campaign for the Homeless). During the second phase, all four evaluations were commissioned by state actors. In the final phase reviewed (2001-2008), two of the nine evaluations were commissioned by NGOs (Focus Ireland/MakeRoom and the National Simon Communities of Ireland).

The authors conclude that both NGO and statutory evaluations have an impact on the policy making process, concluding that “many of the recommendations in the evaluations were accepted and subsequently implemented, whether they origi- nated in reports from the non-profit sector or from evaluations commissioned by government bodies” (p.158). All the NGO-funded evaluations used one or both of the tactics identified above: either contracting authors who can be seen as independent from the commissioning NGOs, or ‘advisory committees’ which serve the same function.

Research as ‘Agenda Setting’.

There are some, mainly faith-based, voluntary organisations providing services to people who are homeless which seek to do so without drawing attention to their work or the lives of the people they serve. However, it would appear that the majority of homeless NGOs wish to draw attention to the extent of homelessness. The motivations for this are numerous and will be discussed later, but at the most fundamental level the organisations are saying: “we are tackling a problem here that, for a variety of reasons, society at large is not paying enough attention to. Pay attention.” 150 European Journal of Homelessness _ Volume 10, No. 3 _ 2016

Research is one attractive and effective instrument for attracting this attention. The publication of research reports is one of the most effective ways to generate media coverage and public debate (Davis, 2009). In this sense, NGOs can be seen as using research to further their attempts at ‘agenda setting’ – that is trying to influence ‘what issues are talked about’ (McCombs and Shaw, 1972).

The emergence of Focus Ireland in 1985 is closely related to an agenda setting piece of research concerning women and homelessness carried out by the organisation’s founder, Sister Stanislaus Kennedy (Kennedy, 1985), a member of the Irish Sisters of Charity. The research came at a time when homelessness in Ireland was perceived to be largely a male problem, and female homelessness was represented only by the image of the ‘eccentric bag lady’ (Focus Ireland, 2011, pp.13-15). The research report, “But Where Can I Go?”, revealed a substantial level of ‘invisible’ female homelessness in Dublin. Furthermore, it identified a significant population of ‘hidden homeless’ women who were not availing of the official shelter system. For the first time, it characterised as ‘homeless’ the women who had been part of Ireland’s Magdalene Laundry system (Smith, 2007) and were now living in convents without tenancy or other rights (Kennedy, 1985). The first services provided by Focus Ireland grew out of a response to these research findings. This set a model of organisational development involving research and service deployment which characterised the next ten years of growth.

Beyond Agenda Setting

But agenda setting is not a value free activity – the issues which are selected and the research question posed have a crucial impact on the how the issue is ‘framed’ when it gets onto the agenda. As has been noted, the selection of the research question is strongly influenced by the objectives of the organisation, and the particular balance of different functions within the organisation at a particular time.

Both O’Sullivan (2008) and Rhodes and Brooke (2010) note the extent to which homeless NGO research included recommendations proposing that statutory organisations produce more accurate data on the extent of homelessness. Since 2014, there has been a substantial improvement in such data and the Department of Housing, Planning, Community and Local Government now publishes the number of individuals using state-funded accommodation on a monthly basis (O’Sullivan, 2016), and the Dublin Regional Homeless Executive publishes ‘info- graphics’ setting out the use of its services. Thus, there is less need for homeless NGOs to commission work on the extent of homelessness. 10th Anniversary Issue 151

The impact of this better access to reliable data has been uneven. Since 2005, the Dublin local authorities have published a twice-yearly point-in-time count of rough sleepers. Even though the NGOs participate in this count, the practice of conducting independent counts persists. The announcement of the ‘official’ figure is frequently preceded or followed by the announcement of ‘unofficial’ counts by one or other of the NGOs with a strong engagement in street work. While the quoted unofficial figure is invariably higher than the official figure, the variance is not sufficiently large to suggest a radically different picture. Nevertheless, the unofficial figure frequently gets more coverage than the official figure, with a regular media implication that the NGO figure is better informed.

This reflects another role which publication of research can have – in particular the publication of a regular measurement of a public phenomenon. It not only fills a gap in knowledge, it also asserts ‘ownership’ of the issue and control over when and how it is discussed. A strange expression of this is the way in which media reports frequently attribute ownership of Government statistics to homeless NGOs, who may be only quoting the official figures to comment on them (for example, Newstalk, 2016; Irish Times, 2016a).

In their eventually successful pursuit of more accurate data, the homeless NGOs were fulfilling one of their traditional tasks – taking on roles which should be the role of the state but which have been neglected. This reflects the role of the NGO as a ‘social innovator’ or ‘pathfinder’, developing areas of social action which the state is too impoverished or hidebound to deliver – and then at some later stage hand them over to the state (National Economic and Social Council, 2005).

Focus Ireland does not see itself as having a responsibility to undertake all the research that is needed, any more than we see ourselves having the overall respon- sibly to end homelessness on our own. In both cases the task must be completed with other civic society actors with the central role being played by the state (through policies that deliver affordable secure housing, access to income, health service etc.). In line with this, we have recommended in our pre-budget submission to Government (Focus Ireland, 2016) that they should ring-fence 0.1% of expenditure on homelessness for research and evaluation – central government funding of homeless services in 2016 was approximately €100m. While this seems a tiny proportion of expenditure when compared with, for instance the research investment in the Canadian Chez-Soi project (Aubry et al., 2015) or the Australian ‘Journeys Home’ project (see Herault and Johnson, this issue) it would represent a massive increase in the research budget in Ireland. Equally we have proposed that the Department of Housing establish a ‘Research Co-ordination Committee’ through which NGOs engaging in research could collaborate to avoid duplication and maximise the impact of our expenditure. At the time of writing, neither proposal has been met with success. 152 European Journal of Homelessness _ Volume 10, No. 3 _ 2016

Research and Service Users

One of the great strengths of homeless NGOs in conducting research is their access to the service users themselves. This can give an insight into the real experiences of people who are homeless which is hard to access through other means. There are real ethical and practical limitations to the extent to which a homeless NGO can see its customer base as a population to be researched. Research can be seen as intrusive and an abuse of the power that the service provider has in relation to the person who is homeless.

One of the strongest proponents of this view in Ireland is Alice Leahy, the CEO of the homeless NGO ‘Trust’. Irish Times (2016b, see also Leahy and Dempsey, 1995) has objected to people who are homeless being used by what she called the ‘research industry’. Her argument is based not only on the level of intrusion (“we must never forget that we are working with human beings, who for the most part have been battered by our society and who for so long have been pushed about as just another number in a cold inhuman bureaucracy.”), but also on the fact that a lot of research has no impact – “we know research is required, but we all know reports are gathering dust all over the place”.

These broad objections to ‘research’ tend to come from homeless NGOs which have a moral stance of ‘asking no questions’ and refusing to ask people anything at all about their circumstances. While these views can be seen as overstated they cannot be dismissed. Given the clandestine lives of some people who are homeless (which may derive from mental health issues or real well-founded concerns about the consequences of the state becoming aware of their whereabouts) such services can reach people who would not engage with services which collect and share information on service interactions. However, this approach tends to go along with a conception of homelessness which is individualistic and a model of service which helps people cope with homelessness while doing nothing to bring it to an end. ‘Accepting people as they are’ can easily result in accepting the world as it is, while lamenting it loudly but ineffectually.

The ethical questions raised by using services data for research have long been understood in the sector, and Focus Ireland established a written research ethics policy over 20 years ago. In recent years, this has been augmented by an independent Research Ethics Committee which can scrutinise relevant research proposals coming either from the organisation or originating from external researchers seeking our cooperation in contacting service users.

One way to avoid intrusion into the privacy of service users can be to use the data which services collect anyway in the course of their work. However, this ‘gold mine’ of data can prove hard to refine – data collected for services purposes often proves 10th Anniversary Issue 153

frustratingly difficult to organised for research purpose. Whenever researchers seek additional data fields, it involves another question to an individual who is in crisis, and diverts the front-line worker from their primary role of supporting the client.

Even in an organisation like Focus Ireland, with a commitment to research going back 30 years, this distance between how services and research staff think runs very deep. Only in the last 12 months has Focus Ireland adopted a practice of requesting permission from customers to allow us to contact them after disengage- ment. Without this it has proven impossible, for data protection reasons, to conduct follow up or longitudinal studies with the customers which we have housed. It took a lot of discussion to build a shared set of priorities with our services so that this permission is now obtained. From next year we will follow up every customers we have housed or prevented from becoming homeless, six months after we have disengaged from them, to ascertain their current status. This will generate not only management information on the effectiveness of our work but also a substantial research base for housing outcomes. While we hope this will produce some insights, we need to remain conscious of the limitations of this data – of course, it only includes people who we already know.

One of the dangers for an NGO which invests in research is that it can become determined to get research corroboration of what it ‘already knows’. While researchers almost always have some expectation of what they will find, some expectation or hypothesis that they are testing, it should deliver some surprises – if it does not then there is probably something wrong with the research programme.

It is self-evident that problems will arise for commissioner and researcher where what the organisation ‘already knows’ is not substantiated by the evidence. How such problems are resolved largely depends upon the contract and integrity of the organisation and the researcher. However, a much more significant problem is nothing to do with integrity, but arises through the unconscious process by which homeless NGOs establish their research question in the first place.

There is an inevitable tendency for an organisation to understand a problem through its own day-to-day experience and the stories related by the people it works with. This is what Sen (1993, 2011) calls ‘positional objectivity’, the process through which an honest and unbiased observer can draw what s/he believes to be an objective conclusion about a situation which turns out to be erroneous because the picture is not fully observed. A judgement can be said to be ‘positionally objective’ if any observer in that position would accept the same judgment.

If you are running a homeless shelter, the majority of your interactions with people who are homeless are likely to be with people who are chronically homeless and have multiple and complex needs. It is reasonable – and appears to be an objective, 154 European Journal of Homelessness _ Volume 10, No. 3 _ 2016

experience-based observation – to say that this is what people who are homeless are like. This perception is reinforced when you have a commitment to listening to the ‘lived experience’ of the people you are supporting. Everything they tell you corrobo- rates this understanding of the nature of homelessness. Further corroboration can then be found when you conduct research which finds a high prevalence of mental health problems, addiction and adverse childhood experiences among the people you are working with. And all this is true. But it turns out not to be the whole story.

A genuine commitment to research must involve a willingness to stand back from the day-to-day reality which dominates your vision and see what the broader picture reveals. Kuhn and Culhane’s (1998) paradigm-shifting research directed attention to a very different picture of homelessness. Their review of administrative data on public shelter use in New York City and Philadelphia provided the insight that people who are chronically homeless represent around only 10% of the total number of people who used public shelters over several years. A further 10% were ‘episodic users’ and 80% only used the shelter for very short periods. The chronically homeless, though only comprising 10% of the individuals who uses the shelters, took up over 50% of the bed-nights and therefore appear to be the authentic face of homelessness.

This dynamic picture of homelessness does not match with the everyday experience of front-line staff dedicated to responding to the problem. The divergence can lead experienced front-line staff and volunteers to be dismissive of all research. The staff of homeless NGOs know for certain what ‘homeless people are like’ because that is what most of the homeless people they engage with do actually look like. It is not surprising perhaps that researchers who contradict what staff and volunteers see with their own eyes are met with cries of “What do these researchers know anyway?!”

Kuhn and Culhane’s insight has become axiomatic in some parts of the homeless sector, but remains totally unknown in others. Homeless NGOs which are not research-informed will still describe ‘the homeless’ as exclusively consisting of people with complex needs and long-term experience of homelessness. Those who are more informed by Kuhn and Culhane’s work will contextualise what they see in their work with an understanding of patterns of inflow and exit from homelessness. In this context, Sen’s ‘positional objectivity’ comes down to not being in a position to tell the difference between ‘stock’ and ‘flow’. If Kuhn and Culhane’s insight can be seen as a genuine paradigm shift in the understanding of homelessness (or allowed to stand as a marker of a broader range of research, including the Housing First research, which resulted in a paradigm shift), the shift is by no means yet complete. There are major organisations, key policy makers and whole systems which continue to operate within the old understanding – and commission research structured in such a way that it appears to confirm the old, static conception. 10th Anniversary Issue 155

Research and Fundraising

A frequent motivation for publishing research is to attract the attention of potential donors. In Ireland, like many other countries, NGOs working with people who are homeless rely to a significant extent on fundraising to provide their services. For instance, 45% of the income of Focus Ireland comes from public fundraising. As a result of this, the fundraising departments of homeless NGOs exert a significant influence on the identity and communications strategy of the organisations. This has a number of potential impacts on the way in which the organisation may wish to portray homelessness, and this may influence any research programme it has.

Government funding in many countries increasingly comes with strings. The constraints which can be imposed by public sector commissioners on service provision have been widely documented, sometimes explicitly shaping only the nature of the service but also applying performance indicators which are so demanding as to exclude certain hard to reach groups. In Ireland, State commissioning of homeless services has not significantly shifted service approaches away from the approaches favoured by the NGOs themselves, and has to date largely proceeded within a shared perspective of what services need to be delivered. Nevertheless, in the broader NGO sector there has been a perception that some state agencies have used the commissioning process to limit the traditional role of NGOs as the ‘voice for the poor’ (Harvey, 2014). Generating a significant proportion of funds from private donations can be seen both as a means of funding innovative front line services and a way of achieving some degree of freedom of expression. However, things are never that simple and the expectations and perceptions of potential donors create a new set of constraints.

To encourage people to donate you need to package homelessness in a way that evokes emotional engagement. Throughout the homeless sector, this consideration has led to a tendency to present the public with the images of homelessness that they are expecting to see, largely involving rough sleepers and street life. As a body of research has emerged which demonstrates that the majority of people who experience homelessness are not chronic rough sleepers (Kuhn and Culhane, 1998; Bejaminsen and Andrade, 2015), there has been increased pressure on homeless NGOs to review the way in which they present homelessness for fundraising purposes. In addition to reinforcing public misconceptions of homelessness, these images serve to shape public policy responses to the problem. A case can be made for a strong interaction between the dominance of the ‘rough sleeper’ image of homelessness and the dominance of shelter provision as the primary response to the problem. 156 European Journal of Homelessness _ Volume 10, No. 3 _ 2016

While homeless NGOs are not the only organisations putting out these stereotypical images of homelessness – even broadsheet media stories concerning homeless families are routinely illustrated with an image of a rough sleeping man – but it is hard for a more complex public perception of homelessness to emerge from the research when the organisations most deeply engaged with the issue use images which reinforce traditional understandings.

The growing proportion of homeless NGO funding which comes from corporate and foundation donors opens up some interesting new opportunities here. Corporate donors bring with them the practice of looking closely at outcomes and are open to arguments about the importance of, say, prevention and long-term sustainable solutions. They are also much more likely to directly support research projects. The last three years have seen Focus Ireland more than double the budget it is able to invest in projects which are broadly defined as ‘research.’ All this increase has come from corporate and foundation support, and includes funding for detailed studies of the impact of homeless prevention programmes, the causes of family homelessness and what can be done to reduce the impact of that experience on children. Such funders also recognise the power of randomised control trials (RCTs), and it is only practical limitations which have prevented such donors funding such research to date.

Research, Timing and People

Both O’Sullivan (2008) and Rhodes and Brooke (2010) show a strong emphasis on research in the early, formative stages of organisations. While the Simon Communities of Ireland have contributed to research over their 40-year history, their most active and influential period of research was in their early years where they produced research which continue to shape policy and legislation today (e.g. Hart, 1978; Collins and McKeown, 1992). Similarly, Focus Ireland followed up the research on homeless women, with significant projects on youth homelessness, care leavers and on standards in emergency shelters (Focus Point and the Eastern Health Board, 1989; Kelleher, 1990; Kelleher et al, 2000; Kelleher et al., 1992), with much lower output from the mid-90s onward.

Both these organisations can be seen to use research to ‘set out their stall’ and mark out the understanding of homelessness which they will address in their services. The predominance of state-funded research from the mid-1990s can be seen as a reflection of the development of state institutions tackling homelessness (Rhodes and Brooke, 2008). It can also be seen as the state and NGO sector playing different roles – the NGO sector having mapped out the main contours of a new understanding of homelessness, with the State sector then responding and filling 10th Anniversary Issue 157

in the key operational elements, while the NGO sector stood back. It is important to note that funding for homelessness increased dramatically during this period, so access to resources was not an issue.

The introduction and continuance of a research tradition is also dependent on the particular influence of individuals. Sr. Stan in the case of Focus Ireland and, in the case of the Simon Communities, their Head of Policy, Brian Harvey – who inciden- tally was one of the founders, and later President of FEANTSA and helped establish the European Observatory on Homelessness. While both Focus Ireland and the Simon Communities continued to commission occasional influential research projects from the mid-1990s until 2008, the emphasis on research of both organisations declined as these individuals became less active or left for other roles. Of the two other organisations noted as making a research contribution by O’Sullivan, Merchants Quay Ireland’s research programme declined from about 2008 when key staff left, while the National Campaign for the Homeless (where the approach had been highly influenced by Harvey) wound up in 1995.

The Focus Ireland research programme has continued and has recently found new energy with adoption of a Strategic Plan which commits the entire organisation to only engage in activities which either serve to prevent homelessness or support sustainable exits from it.

The rapidly changing nature of homelessness in Ireland in recent years has put a high premium on obtaining timely and reliable insights in the housing market. The situation in, for instance, the private rental sector, is changing so quickly that detailed, elaborate research is out-dated before it can be published. Focus Ireland has responded to this by publishing a series of short ‘Insights’ papers, which, for instance, present the detailed trajectories of all families becoming homeless every three months. Much of what is reliably known about the reasons behind the growth of family homelessness over the last three years comes from this Focus Ireland commissioned work. 158 European Journal of Homelessness _ Volume 10, No. 3 _ 2016

Conclusions

Homeless NGOs undertake research for a wide variety of reasons and motivations. In undertaking that research they face a range of methodological, conceptual and data challenges, which can reduce the value and impact of their work. A lot of the work commissioned simply serves to corroborate the scale of the health, mental health and addiction problems faced by people who have spent a long period in homeless services. But it is hard to conclude that these challenges are notably more severe than the challenges facing other individuals or groups carrying out research – or that NGOs have responded less well to these challenges.

In the Irish context, there is a strong case that research undertaken by homeless NGOs has made a substantial contribution at a number of key points in the development of responses to homelessness. In the 1980s and 1990s, NGO commissioned research played a significant role in shifting public policy responses to homelessness away from individualistic, behavioural understandings of homelessness towards structural explanations which ‘locate the reasons for homelessness in social and economic structures…’ (O’Sullivan, 2008, p.21). Simultaneously, consistent pressure from NGOs for better data on the numbers of people experiencing homelessness has resulted in very substantial improvements in data collection and publication, which in turn must feed into better policy making.

As public policy response to these insights, NGO-led research shifted to a different role of attempting to hold the state to its commitments through independent evaluations of progress. These served not only to highlight shortfalls but also successes, and enabled the NGO sector to engage constructively in the review and formulation of strategy in the mid-00s (Brownlee 2008, p.37).

A number of new challenges arise now in setting out a constructive research agenda for NGO commissioned research. Longitudinal research and RCT studies are now recognised as providing the most credible understanding of homelessness and the impact of policies to tackle it, and both raise real challenges for NGOs. Longitudinal studies require a long-term, multi-annual budgetary commitment which is difficult for NGOs to make. Focus Ireland found resources within its own budget to fund the first phase of a substantial research project into the experiences of young people facing homelessness (Mayock et al., 2014), but to carry out the second phase and turn it into a longitudinal study we needed to bring in funding from a number of other homeless NGOs. This form of collaborative funding of research was used in the past (through the National Campaign for the Homeless and MakeRoom) and may be a useful way forward. However, the third phase of the study, scheduled for 2018, is by no means secure. Establishing RCTs for homeless policies raise both logistical and ethical questions which are difficult for NGOs to navigate. 10th Anniversary Issue 159

The decision of an organisation to invest some of its resources in research must reflect a particular pre-existing understanding of homelessness within the organisation (or of an influential person within it). It is appealing to think that organisations which invest in serious research must be predisposed to see homelessness as a problem that can be solved and they are looking to identify what solutions actually work. However, it is equally possible to envisage an organisation committed to, say, providing a soup run, commissioning research to show the beneficial effects of soup or to determine how much soup is required to meet all needs – thus leading to fund raising campaign. There appears to me to be a lot more research which at least aims at ‘problem solving’ than there is ‘soup research’, but this may be due to selective reading.

There remains the deeper question of how homeless NGOs absorb broader research findings and adapt their programmes in response. The research programme associated with the Housing First programme is the most substantial and purposeful strand in homeless research over the last two decades. Many homeless NGOs have now adopted the HF approach, albeit with a wide variation in substantive change or rhetorical adoption. There can be little doubt that a large number of published RTC trials and other detailed analysis is one of the key factors in the widespread adoption of Housing First approaches. However, there can be equally little doubt that this influence did not take place through service managers and policy maker actually reading the published papers – or indeed even the abstracts of them. Most people do not read research papers and few enough people read anything at all after they have completed their formal training – lessons get communicated in other ways, they trickle down, or get brought into the team by younger staff who have just completed formal training. The question is what trickles down and whether, after it has trickled down, it still means the same thing as it did when it started.

For this reason, one of the fundamental challenges for homeless NGOs in relation to research is communicating it to the right people, inside the organisation and in the policy sphere. One of the most important innovations in the Focus Ireland research programme over the last number of years has been our regular ‘lunchtime talks’ and occasional evening talks, both delivered in partnership with Trinity Colleges’ School of Social Work and Social Policy. These talks are open to everyone interested in homeless policy and research in Dublin and regularly attract an audience of around 40 people. Plans in the new year to stream the talks to Focus Ireland services in other cities should increase the audience further. For a number of years we published a regular summary of contemporary relevant research. This was largely superseded by the availability of updates from Canada’s excellent Homeless Hub, but the publications of a localised research summary tailored to the current issues facing front line staff remains a part of our work programme. 160 European Journal of Homelessness _ Volume 10, No. 3 _ 2016

Addressing the question of how to communicate research findings to the staff who interact with people who are homeless in a way which improves practice and outcomes inevitably throws attention back on the most basic of questions – how do we make our research questions relevant? This perhaps is one of the most significant contribution which homeless NGO-led research can bring to the project of research. A constructive and critical dialogue with the front-line staff and service users can lead us to ask the right, timely questions. If we get this right, while standing well enough back to give ourselves the wider perspective, research and learning can contribute to the effective elimination of persistent homelessness as a social phenomenon.

>> References

Aubry, T., Nelson, G. and Tsemberis, S. (2015) Housing First for People with Severe Mental Illness Who Are Homeless: A Review of the Research and Findings from the At Home-Chez Soi Demonstration Project, Canadian Journal of Psychiatry / Revue Canadienne de Psychiatrie 60(11) pp.467-474.

Benjaminsen, L. and Andrade, S.B. (2015) Testing a Typology of Homelessness Across Welfare Regimes: Shelter Use in Denmark and the USA, Housing Studies 30(6) pp.858-876.

Brownlee, A. (2008) The Changing Homeless Policy Landscape: Paradise Lost and Found, in: Downey, D. (Ed) Perspectives on Irish Homelessness: Past, Present and Future, pp.34-42 (Dublin: Homeless Agency).

Collins, B. and McKeown, K. (1992) Referral and Resettlement in the Simon Community (Dublin: Simon Community (National Office)).

Davis, N. (2009) Flat Earth News (London: Vintage Books).

Fitzpatrick, S. and Christian, J. (2006) Comparing Homelessness Research in the US and Britain, European Journal of Housing Policy 6(3) pp.313-333.

Focus Ireland (2011) A Place to Call Home. Twenty-five years of Focus Ireland (Dublin: A&A Farmar). 10th Anniversary Issue 161

Focus Ireland. (2016) Pre-Budget Submission 2017 https://www.focusireland.ie/ wp-content/uploads/2016/09/Focus-Ireland-Pre-Budget-Submission-2017.pdf

Focus Point and the Eastern Health Board (1989) Forgotten Children – Research on Young People who are Homeless in Dublin (Dublin: Focus Point).

Harvey, B. (2014) “Are We Paying for That?” – Government Funding and Social Justice Advocacy (Dublin: The Advocacy Initiative).

Hart, I. (1978) Dublin Simon Community 1971-1976: An Exploration (Dublin: Economic and Social Research Institute).

Head, B. (2008) Three Lenses of Evidence-Based Policy, The Australian Journal of Public Administration 67(1) pp. 1-11

Irish Times, Dec. 2016a http://www.irishtimes.com/news/social-affairs/ peter-mcverry-trust-opens-limerick-social-housing-programme-1.2911337

Irish Times Sep 19, 2016b “Homeless should not be used by ‘research industry’ – campaigner” http://www.irishtimes.com/news/social-affairs/ homeless-should-not-be-used-by-research-industry-campaigner-1.2797364

Jacobs, K., Kemeny, J. and Manzi, T. (1999) The Struggle to Define Homelessness: A Constructivist Approach, in: Hutson, S. and D. Clapham (Eds.) Homelessness: Public Policies and Private Troubles, pp.11-28. (London: Cassell)

Kelleher, C., Kelleher, P. and McCarthy, P. (1992) Patterns of Hostel Use in Dublin and the Implications for Accommodation Provision (Dublin: Focus Point).

Kelleher, P., Kelleher, C. and Corbett, M. (2000) Left Out on Their Own: Young People Leaving Care in Ireland (Dublin: Focus Ireland/Oak Tree Press).

Kelleher, P. (1990) Caught in the Act: The Implementation of the Housing Act 1988 – Housing and Settling Homeless People in Dublin City (Dublin: Focus Point).

Kennedy, S. (1985) But Where Can I Go? Homeless Women in Dublin (Dublin: Arlen House).

Leahy, A. and Dempsey, A. (1995) Not Just a Bed for the Night: The Story of Trust (Dublin: Marino Books).

Mayock, P, Parker, S and Murphy, A. (2014) Young People, Homelessness and Housing Exclusion (Dublin: Focus Ireland). 162 European Journal of Homelessness _ Volume 10, No. 3 _ 2016

McCombs, M. and Shaw, D. (1972) The Agenda Setting Function of Mass Media, Public Opinion Quarterly 36(2) pp.176-187.

National Economic and Social Council (2005) The Developmental Welfare State (Dublin: National Economic and Social Council).

Newstalk.com 14.12.16 http: //www.newstalk.com/reader/47.301/88886/0/

O’Sullivan, E. (2008) Researching Homelessness in Ireland: Explanations, Themes and Approaches, in Downey, D. (Ed) Perspectives on Irish Homelessness: Past, Present and Future, pp.34-42. (Dublin: Homeless Agency).

O’Sullivan, E. (2016) Ending Homelessness in Ireland: Ambition, Adversity, Adaption?, European Journal of Homelessness 10(2) pp.11-39.

Rhodes, M.L. and Brooke, S. (2010) How Does Evaluation ‘Work’? A Case Study of Homeless Services in Ireland, Policy and Society 29(2) pp.149-160.

Royce, D., Thyer, B.A. and Padgett, D.K. (2015) Program Evaluation: An Introduction to an Evidence-based Approach 6th Ed. (Cengage Learning).

Sen, A. (1993) Positional Objectivity, Philosophy and Public Affairs 22(2) pp.126-145

Sen, A. (2011) The Idea of Justice (Cambridge MA: Harvard University Press).

Smith, J.M. (2007) Ireland’s Magdalen Laundries and the Nation’s Architecture of Containment (Notre Dame: University of Notre Dame Press). 10th Anniversary Issue 163

Beyond Homelessness Studies Michele Lancione

Cardiff University, Cardiff, UK

Introduction

Why do scholars research homelessness? It seems to me that asking such a question for the ten-year anniversary of the European Journal of Homelessness (EJH) is not only provocatively interesting, but also needed. I can imagine two possible broad-ranging replies to this query. On the one hand, most researchers do it because they aim for better – i.e., to have a positive impact on the lives of people facing the trauma of continuous displacement. On the other hand, research on homelessness is done mainly because that is what researchers working on homelessness do. That is to say, homelessness is not only what invests people ‘out there’ – the displaced, the marginalised, the poor – but it is also a form of knowledge populated by knowledgeable people who do what they do (surveys, observations, analysis, talks, papers and 10-year special issues) because that is what is expected from them. Following this line of thought, one could argue that researching homelessness – like any other body of knowledge (Foucault, 1990) – is a performance crafted between the will to act and specific institutional schemata, where the latter arguably have the power to affect the former.

Such tension – between one’s own will and one’s institutional status – affects anyone doing research on homelessness, both consciously and unconsciously; the tension is there and cannot be avoided. What can be done, however, is to be reflexive and to learn how it works (see, for instance, The SIGJ2 Writing Collective, 2012). How much of what we do is critical of what we do? How much of it proposes radical alternatives to the canons of research and practice? How much, instead, ends up reproducing the status quo of our – practitioner and researcher alike – professional establishments? These questions are neither new nor simply answered. As the critical turn in disciplines like Human Geography showed, what drives the aim and practice of doing research is never fully questioned, because in questioning it, one must necessarily question the meaning of one’s own profession and status. However, such a questioning is key to social science matters. Only in being

ISSN 2030-2762 / ISSN 2030-3106 online 164 European Journal of Homelessness _ Volume 10, No. 3 _ 2016

open about the limits of what is done and in being honest about the entrenchment of disciplinary knowledge, can renewed, impact-oriented and theoretically relevant research approaches be brought to the fore.

Questioning the relevance of research about homelessness in Europe is quintessential to avoiding that research becoming just a rolling self-referential exercise. This is even truer in current times, when the ground sustaining the bare idea of ‘research on homelessness in Europe’ is shifting at all levels: at the supra-national level, due to the growth of nationalism and sectarian identities, but also at the local level, where the complexification of life at the margins in many of our cities increasingly challenges any definition of what counts as homelessness (Amore et al., 2011). So, if Europe disappears – if not yet nominally, then increasingly so factually – and categories become overthrown by the reality of poverty, of migrant and refugee fluxes, and more, what is left of homelessness studies? Is the field stepping up to the challenges of current times or is it, instead, running behind, trying to catch up? Are, in other words, practitioners and scholars, the readers and makers of the EJH, and the many other organisations involved in the extensive network of homelessness-related ‘stuff’ in Europe doing enough and well enough to rethink the status quo critically and radically?

The straightforward, provocative and partial answer that I would like to give is ‘NO’, we are not doing enough (and perhaps ‘enough’ is not even the right thing to do). I say so not because I believe that the community of homelessness-related activists in Europe is ill-intentioned, lazy or not motivated. Quite the contrary, indeed: motivation for change is there, but the scope of this change is still unclear and the form it should take, undefined. Without assuming any prominence of research over practice, I believe it is fair to say that research is not doing enough to provide if not guidance, at least orientation. This is where I want to address my short commentary: to the need for research to step up to the challenge of the current times; to the need, therefore, for research to be not only practice-oriented in focus (criticising current practices) but to become a driver of new, disruptive modes of being and doing; to the need for research, in a nutshell, to be bold and innovative because this is required by our current condition – by the cuts in welfare provisions (which will translate not only into fewer services, but fewer research opportunities), by the increased precarity of many forms of dwelling (Vasudevan, 2015), by the violence and expansion of continuous forms of displacement (Robinson, 2011; Desmond, 2012), by the fragility of ‘innovative’ policies (Baker and Evans, 2016), and by the already mentioned changes in who makes up the increasing numbers of disenfranchised urbanites (Darling, 2016). 10th Anniversary Issue 165

In the remaining parts of this commentary, I will use the stimulating papers in this special issue of the EJH to sketch a possible direction for the future of homelessness studies and practice in Europe. Such a direction will be – unavoidably for a qualitative, post-structuralist and relatively naïve ethnographer like the author – quite partial and limited. It will also, however, be a starting point, hopefully to generate discussion with, and further provocation from the readers of this journal.

Towards Critical Homelessness Studies?

The European Journal of Homelessness is the golden standard for homelessness study and practice in Europe. Through its pages and its conferences, practitioners and scholars meet to showcase, discuss, provoke, criticise and challenge their actions. The breadth and richness of many of the contributions published in this journal are signs of the interest that homelessness gives rise to in the old continent. That same wealth of knowledge is, however, characterised by a number of worrying limitations. In his beautiful, rich and much-needed paper on the trajectory of homelessness studies in the US and Europe, Nicholas Pleace has correctly pointed out that current research faces a number of risks. He identifies two, with which I very much agree: the assumptive, namely the risk of research taking for granted what homelessness is; and the ‘cultural gravity’, namely the risk of focusing too much on individual experiences of homelessness without linking those back to structural factors and broader contexts. The list could, however, be longer. Although Pleace does not frame it as such, in his contribution he correctly spells out at least one other risk for homelessness research in Europe: the fact that most of it is ‘policy- driven’, related to the evaluation of this or that programme, often on very small scales and with little interaction with broader debates and agendas.

Beside these, three more challenges can be identified. First, there is the risk of homelessness research being done almost in separation from current debates in the social sciences and humanities. It seems to me that research on homelessness in Europe is still too self-referential and out-of-sync with the latest advancements in social thought and theory, and that, arguably, it is currently not able to contribute to those debates in any meaningful manner (Neale, 1997). Secondly, research on homelessness – as I noted in the introduction – still very much takes a responsive, inert approach rather than a proactive one. Despite the engagement that many scholars have with governments and institutions at a variety of levels, the most common modus operandi is that of the consultancy – responding to a pre-determined need – rather than the co-constitution of ideas, agendas and practices. Thirdly, current research seems to follow a positivist quest for quantitative data that seems to lose track of the limitations of big numbers and quantification, perhaps driven by the request from policy-makers for ‘serious’ and ‘reliable’ information. 166 European Journal of Homelessness _ Volume 10, No. 3 _ 2016

These six risks compose a spectrum of challenges that, if taken seriously, could provide food for debate and thought for years to come. How can a field of study overcome future risks of it being assumptive, relativist, policy-driven, self-referential, inert and potentially positivist? Fortunately, the contributions in this special issue offer some reflections on most of these concerns. To me, there are three key suggestions that emerge from this special issue, all of which go in the direction of more self-aware and critical homelessness research in Europe.

The first suggestion comes from the detailed contribution of Lars Benjaminsen and Marcus Knutagård, in which they analyse the cases of Denmark, Finland, Sweden and Norway. Their contribution clearly shows two quintessential aspects that a critical approach to homelessness studies in Europe needs to have at its core. First, and unsurprisingly, the fact that context matters, even more so when one wants to critically assess the national policies and approaches of countries that – for some reason or another – are usually conflated as part of the same cultural-political system. What Benjaminsen and Knutagård have done for the Nordic countries could and should be applied to countries of the east and the south of Europe, which are all too often equated in their pitfalls, while their individual specificities are not sufficiently investigated. Secondly, their contribution shows how structural forces and political orientations lead to very different results in homelessness policies and practices, even in those cases where the context outwardly appears to be the same. The risks of assumption and relativism can, as their contribution shows, be averted through specific attention to contextual dynamics and a critical reading of the nitty-gritty of policy will and policy-making.

The second suggestion concerns a critical approach to data. If, as Pleace reminds us, “gaps in data mean gaps in understanding”, the ‘porosity’ of data itself should never be forgotten. With this term, I mean to highlight the fact that any kind of data is always malleable and prone to instrumentality. This is true both for qualitative data and, despite the general positivist attitude I have mentioned, for quantitative measurements. As Dennis Culhane correctly points out in his contribution, if statistics and fine-grained quantitative data are needed more than ever for better allocation of scarce available resources, those calculations are meaningless – and even dangerous – when undertaken a-critically and when sold as ‘objective’ means to achieving ‘better’ political ends. At the risk of stating the obvious, there is nothing objective about numbers and there is nothing particularly progressive about ‘counting’ homeless people without other, more nuanced aspects being taken into consideration, and these can only be grasped through painstaking qualitative engagements. Moreover, as Herault and Johnson in their contribution demonstrate in their analysis of the Australian ‘Journeys Home’, data is porous also because its effectiveness depends on factors that transcend data itself (like the centrality of policy-makers in the Australian case). The risks of positivism and of being enclosed 10th Anniversary Issue 167

in policy-driven research could therefore be challenged by homelessness scholarship that is clear about the limits and scope of its ‘science’; the science is there, but in terms of practical knowledge rather than techné (Flyvbjerg, 2001; Flyvbjerg et al., 2012).

Last but not least, as an urban ethnographer I can only agree with Pleace when he identifies the need for homelessness scholarship in Europe to engage more (and better) with careful accounts of its people and spaces (ethno-graphy). In that regard, we have arguably a long way to go if we compare our scholarship to the US, but this is a path that, given the aforementioned importance of contextual dynamics and the limits of big data, should be followed as a matter of priority. Good ethnography – namely that which is able to connect local dynamics with broader issues and structures – requires, however, years of effort and engagement. It requires, moreover, an intellectual curiosity and interdisciplinarity that homelessness scholarship in Europe seems to lack. In this regard, the contribution of Vassilis Arapoglou is a welcome attempt to bring the debate on policy mobility, assemblage-of-learning and austerity into the realm of homelessness scholarship. This should be done more and more convincingly in the future to combat the risks of self-referentialism and inertia, as well as to inspire young researchers to go beyond the usual surveys and limited semi-structured interviews, and to engage with in-depth, theoretically- informed and methodologically inspiring fieldwork.

A More Radical Way Forward

The critical approach to homelessness studies arising from the contributions in this special issue would be, to use the terminology from my introduction, ‘enough’ to spark some fruitful debate and to re-orient research (and perhaps practice) priorities. As Pleace notes in his contribution, this is exactly where the EJH is heading with this special issue and with other related efforts. I also state in the introduction, however, that doing ‘enough’ may not even be the right thing to do at this point. This is simply because it may be too late for these debates to be meaningful for the here and now. Radical practice and theory-based changes are needed to engage actively with the scale of current challenges in homelessness in Europe.

The change needed is one that can bring homelessness research closer to its people – whether homeless people, practitioners, activists or others – in ways that are relevant for them and the conditions they experience. Arguably, however, such relevance is not automatically achieved when researchers respond to particular demands – the consultancy and policy-driven approach – but it may come through more elaborate and daring agendas. In what follows, I sketch one possibility. To be clear, and at the risk of repeating myself, the aim of this reflection is not to re-invent 168 European Journal of Homelessness _ Volume 10, No. 3 _ 2016

the wheel of homelessness studies in Europe, but to make it spin in a different direction: one that will hopefully make it more relevant, more open and better equipped to engage with the challenges of the current times. I will offer hints in three chief areas: epistemology, methodology and theory.

Epistemology of practice and engagement If one images a spectrum of possibilities, academic work oscillates between pure theoretical speculation on one end and pure applied research on the other. Arguably, however, it is only at the junction of the two ends where meaningful social science can be produced – one able both to elicit new reflections and inspire new practices. To achieve such an end, however, academics need to step out of their comfort zone and actively pursue meaningful lateral relationships with their non-academic partners. This is, once again, different from a simple contrac- tual agreement, where the layperson contracts the expert scholar to deliver the ‘truth’ about something. As feminist scholarships has pointed out, it is delusional to think that the researcher and the researched are two separate entities that can be maintained as such via the objective means of research (Katz, 1994; Haraway, 1988; Butler, 1999). In reality – as many of us working with practitioners, homeless people and local authorities know – the relationship between ‘us’ and ‘others’ is always constitutive – that is, it always produces relationships of power, knowledge and meaning, even when we do not acknowledge it as being so. As Rose (1997, p.316) puts it:

Following Butler, and Gibson-Graham, there is no clear landscape of social positions to be charted by an all-seeing analyst; neither is there a conscious agent, whether researcher or researched, simply waiting to be reflected in a research project. Instead, researcher, researched and research make each other; research and selves are ‘interactive texts’.

Following this line of thought, it becomes clear that the way we approach the constitutive act of research – i.e., how we go about our epistemology – is quite significant in affecting the meaning and form that our research will take. This could be empowering, horizontal, crafted out of meaningful relationships with our partners; or it could be insignificant – unable to leave any sign of its relevance besides adding another line to one’s publications list. To re-orient homelessness scholarship in Europe toward the first kind of epistemology means to reflect on how we go about doing what we do. It means, in other words, asking ourselves the question with which I started this article: Why do we do what we do? But not only that: it also means acting on that question and its answers, since simple reflexivity won’t go too far in changing established practice. 10th Anniversary Issue 169

One way to go about such active reflection would be to embrace a more activist- oriented approach to research. Such an approach would be oriented both to the production of grounded social theory and to instigating progressive change. Following a recent contribution by Derickson and Routledge (2015, p.6), one could define research-activism as a political ethos guiding ideas and practices that are concerned with an “attempt to find, generate, and resource potential rather than only provide intellectual critique” as well as “to contribute to practices that are aimed at social transformation rather than merely the production of knowledge or the solving of local problems”. Research-activism is not, in this sense, a set of methodologies but an inclination, an epistemology, toward the field and its participants (researchers and researched alike).

The literature provides plenty of examples related to an activist-oriented approach, including Participatory Action Research (PAR) (Mason, 2015) and Solidarity Action Research (SAR) (The Autonomous Geographies Collective, 2010), as well as certain expressions of auto-ethnography from ‘below’ (Reed-Danahay, 1997). These are all activist-oriented approaches, with both limits and opportunities (Pain and Francis, 2003). What unites these approaches is quite simple and it is, once again, more about orientation and sensibility than anything else. As Kindon (2016, p.352) puts it in her basic introduction to Participatory Action Research, research-activism is a matter of ‘attitudes’: from the patronising and compliant “Work with me, I know how to help” to the facilitative, based around an ethos of co-learning, “What does this mean for you? How might we do research together? How can I support you to change your situation?”. Besides the labour needed to actively produce constitutive relationships with our non-academic partners – which requires specific effort from our side to dismantle established modus operandi – research-activism requires open practices of co-working and co-reflection. The researcher moves from being the expert to being a pro-active member of a wider community of change:

These methods and techniques emphasize shared learning (researcher and researched group), shared knowledge, and flexible yet structured collaborative analysis […]. They embody the process of transformative reflexivity in which both researcher and researched group reflect on their (mis)understandings and negotiate the meanings of information generated together (Kindon, 2016, p.355).

Methodologies of creativeness In a recent article (Lancione, 2016b), I argue that in order to move towards meaningful constitutive relationships, academics need to ‘translate’ their research practices and make those open to the ‘other’. Such ‘translation’ is not a one-way process, in which the researcher-translator does all the work (again, in patronising ways); rather, it is a dialogical performance rooted in the ethos of participatory research. In short, translating is a signpost for all those activities needed to make 170 European Journal of Homelessness _ Volume 10, No. 3 _ 2016

research relevant for the people we work with in the here and now. To me, this means both to involve the researched in the makings of the research and, perhaps most importantly, to make the products of research such that they can be used, exploited and turned into relevant artefacts. Translation is relevant for homelessness research in Europe because it can orient scholars to experimentation with new methodologies of engagement and diffusion, bringing them a step closer to the ethos sketched in the previous section.

Writing research papers could be a form of translation, in the sense that their aim is to communicate and engage with broader audiences about a specific set of concerns. However, they only very rarely become translations in the engaging and horizontal sense I have outlined above. Most of the time they are, instead, exclu- sionary and discourage diffusion because of their academic language and the absurdly high prices of many academic publications. If some characteristics of the EJH mark it as very different to the norm (open-source; clear link with various audiences), the translations I am trying to evoke require perhaps further and even more radical modes of engagement. How can we really make our research and its outcomes more open and intelligible?

In my academic work, I argue that the reply to these questions can only be found by bypassing academic practice itself. One way of doing this is through creative methods, namely modes of engagement combining academic, artistic and other non-institutionalised practices. If creative tools have long been used as a means both to collect and analyse field data (Behar and Gordon, 1996), it is only more recently that scholars have been employing them to engage research participants to co-produce and present their findings within and outside academia. A recent example is Marston and De Leeuw’s (2013) collection of works, which shows how it is possible to re-invent qualitative findings in various artistic fashions: from fiction- writing and graphic illustration, to performances, music and installations.

However, these kinds of practices are far from being legitimised. There is a long record of scholars who have tried to cross established boundaries to then be accused of naivety, a lack of objectivity and (quite paradoxically) of turning “away from commitments to engaging ordinary people and offering them voice” (Crang, 2005, p.231). The main issue seems to be that academics have not made enough effort to speak other languages, to cross boundaries and to act in new terrains. This is particularly true when it comes to research on ‘vulnerable’ groups, such as homeless or displaced people. As Cloke et al. (2000, p.147) point out, “[r]esearch and writings on ‘others’ have been produced by, written for and consumed [mainly] by academics”.

Creative methodologies can play an important role both in the production and diffusion of knowledge specific to ‘vulnerable’ people and communities, and in giving ‘something back’ to them and a wider public. They can comprise the doing 10th Anniversary Issue 171

and sharing of narrative writings (Christensen, 2012); ethnographic novels such as the one I wrote to translate my fieldwork with homeless people in Turin, ‘Il numero 1’ (Lancione, 2016b); participatory video-making and documentaries (Sandercock and Attili, 2012; Governa and Puttilli, 2016); participatory mapping and more. The process through which these forms of engagement can be achieved is what I have referred to as translation; it is the creative process through which new, empowering meaning is created.

To be clear, creative translations are not easy and they always involve both collaboration and conflict. Per se, translations are not ‘good’; their meaning depends on how – through which ethos – they are carried out. If they are grounded in a research- activist framework, translations define something very specific and powerful – the re-assembling of research in contextually relevant ways, where the ‘relevant ways’ are defined by the researcher and the people being studied. Translation takes place when researcher, researched and artefacts perform and constitute a productive “coming and going in a borderland zone between different modes of action” (Routledge, 1996, p.406), a zone where the researcher is moved by an ethical and political commitment to bridge the gap with the ‘other’ – to understand their demands and at least partially contribute to their achievement. Homelessness research in Europe could only benefit from playing with creative modes of engagement, not least because of the sheer complexity of the current scenario, which is difficult to grasp by the old methods alone.

Post-categorical theory In a meeting with practitioners and academics that are members of the network Housing First Italia (as I was myself), I mentioned to a colleague that one could write a novel about the efforts required to introduce new policy in such a fragmented context as Italy. By that, as I explained to her, I meant a novel based on long-term ethnographic research, grounded in analysed events, to be done alongside academic papers in order to engage a wider audience and increase impact (as I did with my ethnographic fiction about homelessness). Her dismissive reply is repre- sentative of many I have received – only from academics – since I wrote ‘Il numero 1’: “That would be an easy thing to do”. Other colleagues have told me that fiction writing is not ‘objective’ or ‘serious’, or that it can’t be considered part of ‘our job’ – all claims that had already been dismissed thirty years ago (see Clifford and Marcus, 1986).1

1 For more information on my ethnographic novel, see Lancione, 2016a. For more information about the Housing First Italia network, please refer to Consoli et al., 2016 and, on homelessness in Italy more generally, please refer to the brilliant work of Prof. Santinello and his team of community psychologists in Padova, Italy (Santinello and Gaboardi, 2015). 172 European Journal of Homelessness _ Volume 10, No. 3 _ 2016

Such a reductionist understanding of what an academic can or cannot do is precisely what I am trying to challenge and, I believe, is precisely what a renewed homelessness scholarship in Europe should firmly reject. As intellectuals, we are all called on to contribute to society in relevant and practical ways, which means that, alongside academic publications, we are also called on to experiment with new forms of knowledge-production and diffusion. If our societies and cities change, as they do, it would be foolish to remain anchored to reductionist notions of what is possible and what is not possible within the realm of our jobs.

Reductionism, however, does not apply only to methodology but also affects theory. As I have mentioned, commenting on Pleace’s and Arapoglou’s papers, homelessness studies are currently too bounded and auto-referential. The reason one should reach for more is simple: contamination and assemblage have always been the bread and butter of any serious intellectual project. But how can one make homelessness studies more open to external theories and influences? How could a journal called the ‘European Journal of Homelessness’ aspire to speak to people other than homelessness scholars and practitioners? Before concluding, I launch my final two provocations in this sense.

First, we should confront the problem of ‘defining’ homelessness in the opposite way to what has been done thus far. If homelessness is increasingly harder to define, then we should simply stop defining it. Homelessness should just become synonymous with ‘continuous displacement’ or a form of it (perhaps in the way Pleace (1998) suggested for ‘social exclusion’). In other words, from a bounded taxonomy that defines specific groups, we should move to an open definition focused on processes of experiences, processes of subject formation, and politics (Lancione, 2016a; 2013). The main task would then become to describe those processes in, again, meaningful phronetic ways rather than seeking professional justification and respectability in categorical thinking. To be clearer: if I take myself as example, I would not define myself as a homelessness scholar but as an intellectual interested in matters of continuous displacement – matters that encompass processes that, at the moment, are forcibly enclosed (if accounted for at all) in various ‘typologies’ of homelessness. The mere fact that policies need (for now) those typologies to work should not stop us from rejecting them and, in doing so, starting to re-invent how policy deals with the social issues we are interested in.

Secondly, and even more pragmatically, the EJH could lead to a renewed wave of homelessness scholarship in Europe by, for instance, openly seeking contributions that do not necessarily fit within the canonical remit of the Journal; engaging more with disciplines that have vibrant intellectual communities (such as Anthropology, Geography and Sociology); organising a broad-ranging, cutting-edge and interdisciplinary conference on continuous displacement, whether about eviction, rough 10th Anniversary Issue 173

sleeping, homelessness at home, refugee ‘crises’ etc.; and by encouraging the submission of creative pieces from the arts and humanities and beyond – to name just a few. Most importantly, to rethink the ways in which theory is done within the study of homelessness, the EHJ should continue to be a place where debates take place, where alternative views are welcomed and where experimentation is encouraged. The Journal, and its current Editor, have clearly allowed this to happen in the past ten years; if Europe were to lead a change in homelessness research in the future, that would also be thanks to the work done thus far in these pages.

Openings

Arguing that we need to go ‘beyond homelessness studies’ is a call to revitalise the current state of homelessness studies in Europe and possibly elsewhere. The ‘beyond’ in this sense is not dismissive of what has been done thus far, but a call to make it more current and relevant for our cities and their most disenfranchised populations. The EJH provide the space to continue working on this because of the network they provide and because – as Pleace reminds us – of the advancements made through the ten years of this journal. In this commentary, I have sketched two possible ways to continue and enhance this work. The first is already taking place; it is about making the study of homelessness more self-critical, more reflexive and open to cross-contamination with other disciplines. The second way is harder and more challenging, since it requires deeper changes at the epistemological, methodological and theoretical level – but it is also perhaps more rewarding and meaningful. What unites the two ways is a need for experimentation, criss-crossing and impact in the here and now. In the end, going ‘beyond’ the study of homelessness is not about ‘a’ practice or ‘a’ change, but about exiting the comfort zone of what we do and moving towards, as they boldly used to write and practice some years ago, “un atteggiamento radicalmente critico” – a radically critical attitude (Basaglia, 1968, p.8).

Acknowledgements

I would like to thank Eoin O’Sullivan for the kind invitation to contribute to this special issue and for reading, along with Joe Doherty, a first draft of this paper. All errors (and provocations!) remain mine. 174 European Journal of Homelessness _ Volume 10, No. 3 _ 2016

>> References

Amore, K., Baker, M. and Howden-Chapman, P. (2011) The ETHOS Definition and Classification of Homelessness: An Analysis, European Journal of Homelessness 5(2) pp.19 -37.

Baker, T. and Evans, J. (2016) ‘Housing First’ and the Changing Terrains of Homeless Governance, Geography Compass 10(2) pp.25-41.

Basaglia, F. (Ed.) (1968) L’Istituzione negata. Rapporto da un ospedale psichi- atrico [The Institution Denied. Report from a Psychiatric Hospital]. (Torino: Einaudi).

Behar, R. and Gordon, D.A. (Eds.) (1996) Women Writing Culture (Berkeley: University of California Press).

Butler, J. (1999) Gender Trouble: Feminism and the Subversion of Identity (London: Routledge).

Christensen, J. (2012) Telling Stories: Exploring Research Storytelling as a Meaningful Approach to Knowledge Mobilization with Indigenous Research Collaborators and Diverse Audiences in Community-based Participatory Research, Canadian Geographer 56(2) pp.231-242.

Clifford, J. and Marcus, G.E. (Eds.) (1986) Writing Culture. The Poetics and Politics of Ethnography (Berkeley: University of California Press).

Cloke, P., Cooke, P., Cursons, J., Milbourne, P. and Widdowfield, R. (2000) Ethics, Place and Environment, Reflexivity and Research: Encounters with Homeless People, Philosophy & Geography 3(2) pp.133–154.

Consoli, T., Cortese, C, Molinari, P. and Zenarolla, A. (2016) The Italian Network for Implementing the ‘Housing First’ Approach, European Journal of Homelessness 10(1) pp.83-98.

Crang, M. (2005) Qualitative Methods (Part 3): There is Nothing Outside the Text?, Progress in Human Geography 29(2) pp.225-233.

Darling, J. (2016) Forced Migration and the City: Irregularity, Informality, and the Politics of Presence, Progress in Human Geography. [on-line] Available from: http://phg.sagepub.com/cgi/doi/10.1177/0309132516629004 [17.12.2016].

Derickson, K.D. and Routledge, P. (2015) Resourcing Scholar-Activism: Collaboration, Transformation, and the Production of Knowledge, The Professional Geographer 67(1) pp.1-7. 10th Anniversary Issue 175

Desmond, M. (2012) Eviction and the Reproduction of Urban Poverty, American Journal of Sociology 118(1) pp.88–133.

Flyvbjerg, B. (2001) Making Social Science Matter. Why Social Inquiry Fails and How it Can Succeed Again (Cambridge: Cambridge University Press).

Flyvbjerg, B., Landman, T. and Schram, S. (Eds.) (2012) Real Social Science. Applied Phronesis (New York: Cambridge University Press).

Foucault, M. (1990) The History of Sexuality, Vol. 1: An Introduction (New York: Vintage Books).

Governa, F. and Puttilli, M. (2016) After a Revolution. Public Spaces and Urban Practices in the Core of Tunis, in: M. Lancione (Ed.) Rethinking Marginality. Assemblage, Space and Subjects, pp.42-59. (London: Routledge).

Haraway, D.J. (1988) Situated Knowledges: The Science Question in Feminism and the Privilege of Partial Perspective, Feminist Studies 14(3) pp.575-599.

Katz, C. (1994) Playing the Field: Questions of Fieldwork in Geography, The Professional Geographer 46(1) pp.67-72.

Kindon, S. (2016) Participatory Action Research, in: I. Hay (Ed.) Qualitative Research Methods in Human Geography. Second Edition, pp.350-370. (Melbourne: Oxford University Press).

Lancione, M. (2013) Homeless People and the City of Abstract Machines. Assemblage Thinking and the Performative Approach to Homelessness, Area 45(3) pp.358-364.

Lancione, M. (Ed.) (2016a) Rethinking Life at the Margins. The Assemblage of Contexts, Subjects and Politics (London: Routledge).

Lancione, M. (2016b) The Ethnographic Novel as Activist Mode of Existence: Translating the Field with Homeless People and Beyond, Social & Cultural Geography [on-line] Available from: http://dx.doi.org/10.1080/14649365.2016.1231 336 [17.12.2016].

Marston, S.A. and De Leeuw, S. (2013) Creativity and Geography: Toward a Politicized Intervention, Geographical Review 103(2) pp.iii-xxvi.

Mason, K. (2015) Participatory Action Research: Coproduction, Governance and Care, Geography Compass 9(9) pp.497-507.

Neale, J. (1997) Homelessness and Theory Reconsidered, Housing Studies 12(1) pp.47-61. 176 European Journal of Homelessness _ Volume 10, No. 3 _ 2016

Pain, R. and Francis, P. (2003) Reflections on Participatory Research, Area 35(1) pp.46-54.

Pleace, N. (1998) Single Homelessness as Social Exclusion: The Unique and the Extreme, Social Policy and Administration 32(1) pp.46-59.

Reed-Danahay, D. (Ed.) (1997) Auto/Ethnography: Rewriting the Self and the Social (Oxford: Berg).

Robinson, C. (2011) Beside One’s Self. Homelessness Felt and Lived (New York: Syracuse University Press).

Rose, G. (1997) Situating Knowledges: Positionality, Reflexivities and Other Tactics, Progress in Human Geography 21(3) pp.305-320.

Routledge, P. (1996) The Third Space as Critical Engagement, Antipode 28(4) pp.399-419.

Sandercock, L. and Attili, G. (2012) Unsettling a Settler Society: Film, Phronesis and Collaborative Planning in Small-town Canada, in: B. Flyvbjerg, T. Landman and S. Schram (Eds.) Real Social Science. Applied Phronesis (New York: Cambridge University Press).

Santinello, M. and Gaboardi, M. (2015) Presentazione del numero. Marginalità estreme [Presenting the Numbers: Extreme Marginality], Psicologia di Comunità 2 pp.5 -7.

The Autonomous Geographies Collective (2010) Beyond Scholar Activism: Making Strategic Interventions Inside and Outside the Neoliberal University, Acme 9(2) pp.245-275.

The SIGJ2 Writing Collective (2012) What Can We Do? The Challenge of Being New Academics in Neoliberal Universities, Antipode 44(4) pp.1055-1058.

Vasudevan, A. (2015) The Makeshift City: Towards a Global Geography of Squatting, Progress in Human Geography 39(3) pp.338-359. This publication has received financial support from the European Union Programme for Employment and Social Innovation “EaSi” (2014-2020)

The information contained in this publication does not automatically reflect the official position of the European Commission European Observatory on Homelessness European Observatory on Homelessness

European Journal European Journal of Homelessness Homelessness of Homelessness The European Journal of Homelessness provides a critical analysis of policy and practice on homelessness in Europe for policy makers, practitioners, researchers and academics. The aim is to stimulate debate on homelessness and housing exclusion at the Journalof

European level and to facilitate the development of a stronger evidential base for policy development and innovation. The journal seeks to give international exposure to significant national, regional and local developments and to provide a forum for comparative analysis of policy and practice in preventing and tackling home-

lessness in Europe. The journal will also assess the lessons for European Europe which can be derived from policy, practice and research from elsewhere.

FEANTSA works with the European Commission, the contracting authority for the four-year partnership agreement under which this publication has received funding.

The information contained in this publication does not necessarily reflect the position or opinion of the European Commission.

194, Chaussée de Louvain n 1210 Brussels n Belgium Tel.: + 32 2 538 66 69 n Fax: + 32 2 539 41 74 Issue Anniversary th [email protected] n www.feantsaresearch.org 10 Volume 10, No. 3Volume 10, _ 2016